Latest Posts

The Reliants Project in Kumu

Finally, I’ve managed to embed an anonymised version of the latest Reliants Project network graph onto the front page of the site. I think it is much more interesting to engage with an interactive map than static images. Back in 2017, I shared maps showing my London personal network before and after pairing up with my partner. Before that, I shared a global map showing how my network has evolved over my adult life. Over the last year, I have introduced my partner to many people within my global network. We have also introduced many of our friends to each other. The resulting 2018 graph is more complex than the previous versions and shows a more developed London network.

Reading the map. In Kumu, nodes are called elements and edges are connections. Each element represents an individual in my personal network. The connections show who knows who in that network. In this map, the colors indicate which geographical group the individual is part of (United States, Hong Kong or the United Kingdom). The large elements identify the ‘reliants’ (close friends and family that I rely on) in my network, which happen to be evenly distributed across the 3 geographies. There are a handful of characteristics that describe each element, including a ‘tag’ that indicates the context in which we met (public event, school, work, introduced, etc). For now, I’ll focus on a few of the key features that I’ve been playing with:

  • Focus feature
  • Community Detection
  • Highlighting elements and connections
  • Filtering elements and connections
  • Social Network Analysis

Focus feature. One of the ways you can interact with this embedded map is using the focus feature. You can select an element on the map and then click the sight icon on the right-hand-side of the window, below the zoom icons. This makes the selected element the focus and displays all the direct connections of this individual. You can then select the up arrow to expand the focus to indirect connections. This gives you a sense of how many degrees of separation there are between different elements based on where they sit on the map.

Community detection. You may be able to see clusters of elements within the map. These are referred to as communities in Kumu and represent groups of individuals who know each other. They might be work colleagues or friends from the same school. In some cases they are easy to identify visually, but Kumu also has a tool that helps you to identify them mathematically. It was able to identify 7 distinct communities within this map which turned out to accurately reflect how I would categorise my friend groups.

Highlighting and filtering. The giphy below runs through the different views of my network map, which take advantage of the highlighting and filtering functionality: 1. Default view of the whole network, 2. Highlighting my HK network, 3. Highlighting my US network, 4. Highlighting my UK network, 5. View with elements ‘pinned’ and introductions I’ve made filtered out.

via GIPHY

Social Network Analysis. Kumu also allows you to generate a handful of metrics. Unsurprisingly, the individuals that rank high on most metrics are also reliants. This accurately reflects my behaviour, since I make an effort to introduce my family, partner and close friends to my network. Of course, the metrics are only as good as the data. Since I built this data set by hand and may be missing some connections, the results are likely to be imperfect. Also, it’s a reflection of the network from my point of view and doesn’t consider the strength of connections between these elements.

That’s all for now! I’m working on a handful of different graphs and look forward to sharing them on the blog soon. As usual, I welcome any feedback or suggestions!

Continuous participatory change

This morning I’m sitting on the tiny, south-facing balcony of our flat with a coffee-filled mug with red letters that say “niet normaal”. Even though the balcony is overwhelmed by vines, I somehow haven’t managed to keep our mint alive. This is the third attempt to complete this update, trying to find a narrative thread through the last 12 months. While the phrase “continuous participatory change” is used as an approach to organisational transformation, I think it a powerful way to approach all of life. For most, change is difficult, uncomfortable and disorienting. More than ever, it’s also the only constant. Over the last year, I have been learning new ways to thrive in these conditions and navigate change.

A habit I adopted last year is to ask myself questions from Changing on the Job, such as:

“What assumptions about the world underpin my or others actions and opinions?”

It has helped me reframe many challenging situations and identify some of my own blind spots. For example, I strongly believe that creating more connections within a network makes it more resilient and that the network effects benefit everyone involved. I don’t make a habit of asking for permission to make these connections from anyone. However, I have come to appreciate that not everyone is comfortable with this and some perceive my actions as an imposition or even a threat. Finding common ground through understanding intent has helped to bridge these types of gaps.

In January, my team was given the remit to organise ourselves. It’s part of a larger effort by Anthemis’ founders to resist growing the company using old organisational models. I won’t sugar coat it, the shift has been tough for many. Some of the many tools we have built habits around are the OS Canvasretrospectives, and integrative decision making. Building a team that learns together has been incredibly gratifying. We are quick to create experiments we want to try and shed processes that don’t serve us. If anyone is interested, happy to share more details about our journey.

Another set of tools I’m exploring to help navigate change are mental models. Few of them are used outside of their domain, and mostly for the purpose of making investment decisions. This spring I learned 113 of the most common models using method of loci and spaced repetition techniques. What does this look like practically? Imagine you pictured a baby with little devil horns overlaid on the steps of the Taj Mahal and immediately thought, “of course, regression to the mean – the 13th numeracy mental model.” It’s amazing how easy it is to recall information that you learn this way. I find myself noticing references to and uses of mental models on a daily basis now that my mind has been primed to spot them. If you want to get a taste, you can try it out here.

 

How I invested 8760 hours in 2017

Inspired by Kunal Gupta’s How I Invested 2504 Hours post, I decided to perform my own audit for 2017. My goal was to compare how I actually invested my time to the ideal 24 hours I imagined back in 2016. This way, I could decide if I wanted to change my time allocation for 2018 and implement habits to help me do so. Back in 2016, here’s what I outlined for a typical day:

pie

  • Sleep and restorative (sleep and naps) – 7 hrs / 30%
  • Creative and productive (content creation, workshops, culture) – 7 hrs / 30%
  • Eating and social (meals, coffee, drinks) – 5 hrs / 20%
  • Active and physical (exercise, walking, sex) – 3 hrs / 12%
  • Quiet and reflective (meditation, reading, bath) – 2 hrs / 8%

 

The results. All in, I was able to account for about 80% of my time in 2017. Of that time, a third was scheduled, a third was unscheduled estimates and a third was sleep. While the math is obvious, it was still surprising to internalise that spending 85 hours on something was only the equivalent of 1% of my time. Similarly, it was helpful to remind myself that if you work 8 hours a day and sleep 8 hours, you still have 8 more hours to invest. Here’s roughly how my time broke down:

  • Sleep and restorative (sleep and naps) – 8 hrs / 33%
  • Creative and productive (workshops, meetings, creative time, cultural activities) – 6 hrs / 25%
  • Eating and social (meals, coffee, drinks) – 2.75 hrs / 12%
  • Active and physical (exercise, walking, sex) – 1.25 hrs / 5%
  • Quiet and reflective (meditation, reading, bath) – 3/4 hr / 3%
  • Unaccounted for – 5.25 hrs / 22%

Reflecting on this, I suspect that the 5% gap between my ideal and reported creative time is predominantly due to unscheduled time at work and on weekends for personal projects. As for the 8% discrepancy in eating and social time, I am sure a large chunk of that unscheduled time was used to build a relationship with my partner. It’s easy to believe that a portion of that time was also used to be quiet and reflective.

The 7% difference in active time cannot easily be rationalised, as I doubt much of the unaccounted time could be described this way. If there is one thing I’d like to change in 2018, it’s to spend more time each day being physically active. I’ve already identified a stretching routine that I’d like to incorporate into my daily habits.

Although I haven’t done this analysis yet, I suspect that portion of time I spent building new relationships outweighs the time I spent nurturing existing relationships. This is to be expected, since I spent a fair amount of time meeting my partner’s friends and family. However, I’d like to make sure that I make more time for current friends in 2018.

Since my workday was largely accounted for, I was able glean a fair number of insights. I was able to determine how much of my time was focused on core vs. administrative activities (roughly 85/15). The analysis also gave me a sense of how much time I spent developing myself and others (about 10%), participating in events (another 10%), and on recruitment (5%). Once I have finalised my goals for 2018, this will help me figure out what I should dial up or down.

A few random stats that surprised me:

  • Scheduled 128 coffees, teas or drinks with others
  • Held 92 interviews
  • Attended 59 events (conferences, talks, performances)
  • Took 19 flights
  • Read 17 books

Of course, some activities can’t be neatly categorised and I often do two things at once. The epitome of efficiency is walking to work while taking a conference call (physical activity, commuting and productive time all in one)! Also, it was near impossible to categorise my invested time into the buckets of foundation habits, experiences & relationships, and impact. Will think about a clever way to do that for 2018.

The how. You’d be surprised at how much data you can collect for this. Google calendar and other apps, such as Strava, are a great source of information. You can even download apps that will tell you how much time you spend doing various activities online. I also estimated some of the data, as I do not have a sleep tracker and I don’t make note of every time I meditate or walk to work.

Work: After exporting all my calendar data, I used Google Sheets to organise my scheduled activities into categories such as workshops, development, interviews, recurring, events, etc. This was the most manual part of my analysis. This helped me to understand what percentage of my time I was spending on the business vs. in the business.

Meals: Since I’m fairly rigorous about how I add calendar invites, I was able to search my Google Calendar data for breakfast, brunch, lunch, dinner, coffee and drinks. I then estimated 30 minutes each for all meals I ate by myself or unscheduled with my partner (I almost never skip meals).

Travel: In this case, I estimated my regular commute. I was able to pull my flights directly from Google Calendar and add up the flight times and estimate the travel to and from the airport. I handled trains in the same way. If Uber is your thing, you can even export your trips to csv. Turns out I only spent 16 hours in an Uber that I hired this year. My partner booked about half the time and he purchased a car in July.

Reading: In order to estimate this, I used my Goodreads list to identify books I read this year and how many pages they have. Then I multiplied the entire number of pages by a reading speed of 2 minutes per page. I also exported my Pocket article list and multiplied the number of articles read by 5 minutes, which is the average length of a Medium post.

Sleep: Since I’m very protective of my sleep, I estimated an average of 8 hours a night for the entire year.

With some foresight, it should be much easier to track. I’m going to attempt tagging items in my calendar to make this process easier come next December. I might also look at tracking specific activities that I want to improve (such as physical activity).

Let me know if you’ve ever done something like this and have any tips to share!

 

Primary relationships impact personal network structures

In my post “Change over time“, I hypothesised that coupling up and separating with someone would have a significant impact on the structure of an individual’s personal network. After my divorce many years ago, my network structure shifted from one that was compartmentalised to that of a ‘sampler’. Little did I know that I would soon have the opportunity to actually test this theory.

Between 2015 and 2016, I tracked how my London social network grew from a small group of pre-existing connections to a reasonably strong support system. In the resulting graph, I emphasised the role that non-local contacts had in helping me expand my community. Now I’ve taken 2016 data and compared it to the present in a new visual:

2016-single-2017-couple

The most dramatic differences between the 2016 and 2017 visualisations are the new node sitting at the centre of the graph and the cluster of new nodes on the far right. That central node is my new partner and the cluster to the right is the portion of his London network that he has introduced to me. His dominant London network structure is that of a ‘tight-knitter’, where most of his local friends have all known each other for a long time.

On the left-hand-side of the 2017 graph, you can see that I continued to bring the network together through introductions. The nodes are more interconnected and resulting graph denser. You can also see the introductions I’ve made between that network and my new partner. The cluster of people on the left-hand-side that already knew each other are predominantly work friends.

Over time, I anticipate that the two clusters on either side of my partner will be linked together through our introductions. Whether those introductions form lasting bonds will impact how this joint structure evolves. I will track the data in order to create a time series similar to my previous posts showing how the structure changes over time.

Honeycomb hexayurt

The first time I went to Burning Man, my experience was made much more comfortable by fellow campers who had prepared temporary shelters and offered me a spare. Insulation from cold, reflection of sun, firm walls, and being able to stand up removed many of the common irritations of camping. When our group began planning for this year’s trip to the Burn, I wanted to give them the same level of comfort. I set out to design a structure that would provide all of those benefits and some privacy, while minimising the amount of construction materials we’d have to buy and transport.

The design I arrived at is based on the 6′ hexayurt, one of the many designs by Vinay Gupta. In order to reduce the amount of building materials required, I arranged them in a honeycomb layout. This saved us 4 sheets of 4’x8′ insulating foam and about a roll of filament tape. It could easily be expanded to incorporate more hexayurts, seemingly indefinitely (though it does make sealing much more difficult). Here’s the recipe.

hexayurt-design-revPreparation materials:

  • 18 | 4’x8′ sheets of 1″ thick polyisocyanurate insulating foamboard, such as Rmax
  • 4 | 50 yard roles of multipurpose foil tape at least 1.75″ thick, such as Nashua
  • 1 | utility knife with at least 6 blades
  • 1 | 4′ T-square
  • Something long and flat to kneel on and distribute weight

We prepared the materials (cutting structural parts, cutting out doors, covering the raw edges in foil) before Burning Man. We followed the schematic in the image above to make A and B parts. I won’t get into the play-by-play, but we learned 3 useful things:

  • A 4′ metal T-square significantly speeds up the cutting process
  • Marking the matching doors and door frames saves time later
  • All of the finished parts will actually fit into a Dodge Grand Caravan

shelter4

When we arrived on the playa, we pulled out the rest of the materials and set to work.

Assembly materials:

  • 3 | 60 yard roles of 3′ bi-directional filament tape, such as this
  • 8 | rebar stakes, like these
  • 8 |tennis balls (with slot cut to fit over rebar stakes)
  • 4 | waterproof tarps, at least 7’x8′ each to place inside the shelters (also great for carrying the materials)
  • 150 feet of nylon cord, for example 3 of these

shelter3Most of the time and frustration-saving tips became relevant during assembly. I cannot stress enough how valuable it was to first use small pieces of tape to hold the structure together before using long, continuous pieces of tape to seal the structure. This way we could modify the spacing and readjust if the pieces didn’t line up quite right on the first try. Also, if you happen to have a friend that is 6’7″, bring them. We assembled the parts in this order:

  • Assembled all 4 roofs, made of 6 triangles each
  • Put together 2 sets of walls, made of 5 squares and 1 door frame each (shown in blue and marked 1 on the diagram)
  • Arranged the 2 sets of ‘blue’ walls 4′ apart (with door frames facing away from each other) and added the spacer wall (shown in red and marked 2 on the diagram)
  • Assembled 2 sets of walls, made of 2 squares and 1 door frame each (shown in green and marked 3 on the diagram)
  • Attached the last 2 sets of ‘green’ walls to the structure
  • Attached the 4 roofs on top
  • Sealed the completed structure with filament tape (the longest, continuous pieces of tape manageable)
  • Made 4 sets of tape hinges for the doors
  • Tied 4 loops of nylon cord together to make a crown, which we then attached to rebar stakes in the ground

And here’s the result, our very own honeycomb hexayurt village:

OLYMPUS DIGITAL CAMERA

OLYMPUS DIGITAL CAMERA

The list of potential failure points makes me cringe even now:

  • Designing the shelter accurately
  • Estimating material quantities correctly and availability of material stock
  • Coordinating timely material delivery
  • Preparing the materials accurately and without wastage
  • Fitting the prepared materials into the Caravan and RV (our methods of transport)
  • Ensuring the materials were dropped off at our camp
  • Assembling for the first time in the dark (with no prior experience)

Anyone considering replicating this design might want to do a better job of reducing the possible points of failure. Nonetheless, during one of the hottest Burns on record, we stayed surprisingly cool!

 

 

Standing the test of time

It’s been over 9 months since I adopted 3 ‘networked’ habits to test if I could make them more resilient than stacked habits in the face of change. This is one of the four intentions that I set for myself during my 2016 yearly review. Since then I have:

  • Transitioned from being single to having a partner
  • Moved from East London to North London
  • Managed a new commercial partnership
  • Traveled a total of 8 weeks

In the past, any one of those situations would have thrown me off the wagon. Changes to workload, routine and certainly life stage have historically had a big impact on my foundation habits. However, this time I’m thrilled to report that I am still consistently eating healthy (38 AmazonFresh orders placed in 2017), walking (at least 1:30 hour ~4x/week), and meditating (~4x/week). The walking is either to commute, attend meetings or even as a meeting itself).

Of course, I’m still linking the new activity to another one that was already deeply embedded. The difference is that I’m not layering other new activities on top precariously. There is a one-to-one relationship between the habits that I have consistently done for years and the newer activities that I hope to take root.

Also, I’m being much more realistic about what is sustainable, where I might have historically had a much more ‘all or nothing’ attitude. If I am unwilling to meditate for 2 hours like the vipassana guru recommends, at least I can consistently achieve 15 minutes. If I won’t run 5km every morning, at least I can embed 6hrs of walking into my week. Perhaps eating spinach soup for lunch every day is unreasonable, but following Michael Pollan’s food rules certainly is.

The more recent habits are also implemented in a way that reinforces them in a ‘networked’ way. Having healthy food in the house means that my partner and I cook together more than we might otherwise. It also saves me money on lunches out. Walking frequently allows me to think about and reflect on my work. Walking meetings change the nature of the conversations I have with others. These additional benefits to maintaining these activities also help to reinforce those behaviours.

Here’s what is working for me:

  • Thinking carefully about how to design the new activity to make it reinforcing (i.e. combine it with other activities I enjoy and value, such as spending time with people I care about).
  • Making sure the old habit I link to is deeply rooted and not context specific (e.g. I always set an alarm regardless of where I am).
  • Having clarity about how I prioritise these activities. Sufficient sleep is more important than walking, eating a healthy breakfast is more important than meditating.
  • Being aware when the old habit is uprooted and making adjustments accordingly (e.g. changes to travel patterns).

Let me know what works for you!

 

 

 

 

 

Eight weeks of ‘networked’ habits

Gotta celebrate the little wins! At the end of last year I completed another year in review and identified 3 areas that I wanted to focus on for 2017. One of areas of focus is intensity, where I’ve asked the question:

How can I make my foundation habits more resilient to changes in my life’s intensity and rhythm?

Limitations of habit stacking. I’ve had the ongoing issue of stacking habits only to watch them fall like a house of cards in the face of change. When I fall off the wagon, it can take me months to get back on again. The only foundation habit I consistently maintained in 2016 was sleeping 8 hours a night. In Q1 I kept an incredibly aggressive physical training schedule. In Q2 I ate healthy and cooked consistently. Most of Q4 I meditated regularly. There was no stretch of time when I managed to consistently incorporate all 4 into my life together.

Enter networked habits. Yeah, I know, if your tool is a hammer everything looks like a nail. However, I do find that my most resilient habits are connected to several other behaviours, habits and or beliefs that help them become more entrenched in my life. Not in a ‘stacked’ way (one after another), but in a ‘networked’ way (nodes with several interconnected edges). They still require a trigger habit, but whenever possible that habit is already deeply rooted. Here they are, in all their simple glory:

1 | Turn off alarm > open timer app. This simple habit has helped me consistently meditate for the last 8 weeks. Usually I set a 15-30 minute timer for my meditations, but I often got distracted by notifications and wasted that valuable time. Now I turn my phone to airplane mode when I go to sleep (which silences notifications from people in different time zones) and immediately set the timer for meditation when I wake up. Since I set an alarm almost every day, I’m hoping this habit will stand the test of time.

2 | Receive Amazon Fresh delivery > order Amazon Fresh delivery. Trips to the market are first to become collateral damage when workloads increase and schedules become intense. I have a hard time justifying a late night trip to Tesco, in part because their protein selection is embarrassingly poor for someone who doesn’t eat meat. Also, I much prefer to support small, local businesses and artisanal purveyors. Amazon Fresh has become a saviour in this regard. I can still buy produce from Chegworth Valley, salmon from Forman and Field, and bread from Gail’s within minutes. This means I always have healthy food around for simple cooking. This also unlocks the opportunity for more dinner parties and brunches, which I love to host.

3 | Add event to calendar > ask ‘can I walk there?’ Two heels are better than four wheels. Or any form of public transportation. I have made a habit of asking myself ‘can I walk there?’ when I add plans to my calendar, which ensures that I build in the time to do so. It’s had a number of wonderful, unanticipated side effects above and beyond incorporating physical activity into my day. This time has become incredibly valuable for thinking about new ideas and through problems. It also gives me the opportunity to explore new neighbourhoods and feel like a tourist in my own city.

Will keep you posted as to whether these are also vulnerable to ‘wagon-falling-off’.

Celebrating friendships

After all the effort of building a dataset for The Reliants Project, it’s been great the reap the rewards with countless ways to explore and visualise the data. I’ve decided to focus first on the reliants, my closest relationships.

To give readers a sense of the breadth of the group categorised as reliants, here are some reference points. They include family members I’ve known since birth, others I’ve built relationships with spanning 25+ years, as well as people I’ve gotten to know within the last year. I met an equal number through direct introductions and public events and there are even two that I met serendipitously. They are overwhelmingly male, but very diverse in terms of nationality and ethnicity. Their ages span from mid-twenties to retired, though the majority are 25-45. Almost all of them have moved internationally and have lived in the same city as me at some point, though there are a couple of exceptions. Roughly half of them are married and/or have children, however few had reached this life stage when I met them. Beyond family (2 people), two pairs have relationships with each other that pre-date me. Even though all my reliants have been incredibly important to me at various stages of my life, there has been a 66% turnover in my reliants over the last 15 years.

reliants-relianttimeseries

One of the first things that struck me after creating the visual above is the consistency of node quantity overall (11-15) and the balance of nodes between locations, which is evenly distributed (4-7). Unsurprisingly, this closely reflects the Dunbar “rule of three“. It’s as if my brain makes space for a certain number of very close relationships and adjusts as life events and stage changes occur. Also, my network has become increasingly interconnected over time, as I get the opportunity to introduce reliants to each other in an effort to make those relationships more resilient. I suspect that trend will play out as my time in London lengthens.

In the series of graphs below, I take each of the reliants depicted in 2015 (above) and visualise them in the context of our shared network. The sequence is based on the order in which I met each of them. You can see how 1, 6, and 13 straddle locations, either between US and Hong Kong or Hong Kong and London. You can also see how close the relationships are between 7 and 8 as well as 9 and 10 because of the consistency of the network visuals and their position within them. It’s also easy to spot the people who have never lived in the same city as I have, such as 14. All of the graphs of my new London reliants look very similar to 15. The combination is a lovely fireworks display!

reliants-fireworks

 

One of the many paths that I would like to explore is qualifying both the nodes and edges (reliants and our relationships with each other). While there are many ways to assess the nodes (such as the Social Style Model I used in my questionnaire), there are few ways to assess the edges. Recently I had the pleasure of meeting someone very familiar with the Relational Proximity Framework, which acknowledges the multiplexity of relationships and seeks to assess them using 5 domains: power, information, communication, purpose and story. Hoping that I can find an elegant way to display those dimensions and add richness to the data visualisations and insights that can be generated from them.

As always, would love any suggestions!

Change over time

If you’ve met me, you’ve probably heard me say “change is the only constant” more than a few times. It’s been incredible to reflect on how much my personal network has evolved since university. After 12 months of The Reliants Project focused on my new London network, I decided to shift focus to 3 areas inspired by that exploratory research:

  • Building a more complex visual of my entire personal network in the hopes that it will give me a more accurate representation of change in my network over time
  • Visualising how new connections transition between the categories of stranger, acquaintance, friend and reliant (and even loosing touch) over time
  • Identifying how significant life events (e.g. moving, marriage, parenthood, divorce, career shifts) impact connections’ positions within the network

While I gathered data over the last 15 years, it was hard to reach back beyond 2004 (introduction of Gmail) with much accuracy. Nonetheless, this data captures 2 international moves (Massachusetts to Hong Kong in 2008, Hong Kong to London in 2015), my divorce (2010), and 4 career shifts. The first time series visual I created based on that data is below (click to enlarge).  If you’re interested in more detail about how it was made, you’ll find it at the bottom of this post.

reliants-timeseries-location

Emerging patterns and hypotheses. It’s still early days, but I was surprised to see how compartmentalised my social groups were during my married years. I suspect that pairing up and separating can have a significant impact on someone’s personal network structure, specially if the way they typically structure their network is different from their partner (one tight knit group vs. a diverse set of one-to-one friendships).

Immersion, by some clever folks at the MITimmersion_boston_categories Media Lab, also helped me see these patterns. It visualises a network based on email metadata and can display any period of time you wish. It was interesting to see how my email graphs reflected the Polinode graphs I made with my own dataset. For example, up to 2008 I was living in Boston and had 3 main social groups outside my family (local Boston friends, designers and high school friends). They show up clearly in both the visualisations.

Basic personal network stats: 10% reliants (including family), 60% friends, 20% acquaintances (keep in mind that most acquaintances were added because they introduced me to someone who became either a friend or reliant at some point), and 10% lost touch. The proportion of reliants is in line with my London network.

How did I meet them? 45% through introductions, 30% at public events or ‘in between’ places, 10% at private events, and 15% through school or work. Half of the friends I’ve made in London are also as a result of introductions.

Survey. In order to add more richness to the data, I asked the network to complete a short survey focused on a few key aspects of their social characteristics using Typeform. They made an educated guess regarding their current social network structure (based on recently published research by Janice McCabe),where they sit within their family (an indication of their first personal network), their social type (based on the TRACOM Group Social Style Model), and any life changes that may have impacted their personal network over the last 15 years. About 60% of the network responded to the survey. Below are the actual survey questions and results.

1 | Based on your current friendships, would you describe yourself as a ‘Tight-knitter’, ‘Compartmentalizer’ or a ‘Sampler’?

Tight-knitter: You have one dense group of friends in which nearly everybody knows one another.
Compartmentalizer: You have two to four “clusters” of friends who don’t know each other; one “cluster” may comprise people you have fun with, while another could be made up of people whom you turn to for work-related support or advice.
Sampler: You have one-on-one friendships, rather than groups of friends, and don’t rely on friendships for a sense of belonging.

survey-structure

2 | Within your family, which child are you?

survey-child

3 | Any guess what your Social Style Model type is?

Analytical: control their emotions but tend to ask questions rather than tell people what to do. They are focused on accuracy, and they act deliberately to achieve that end. Others see them as slow-paced and detail-oriented.
Amiable: show their emotions openly and prefer to ask questions rather than tell others what to do. Relationships, feelings and personal security are important to Amiable Style people. Others see them as friendly and warm.
Expressive: show their emotions and speak assertively. They enjoy sharing their ideas and perspectives openly with others. Others see them as creative, but unfocused.
Driving: control their emotions and speak assertively. They prefer to control a situation and are focused on the big-picture. They are often seen by others as highly efficient and not concerned about relationships or feelings.

survey-social-type

4 | Based on academic research, experiences of moving, marriage, parenthood and divorce typically have the most impact on individual’s personal network. Have you experienced any of these significant life changes in the last 15 years?

survey-change

5 | Out of curiosity, is there anyone that I’ve introduced you to that has become a friend, close friend or reliant? If so, who? Just to clarify, for the purpose of this research I categorise friends like this:

Friends: would invite to a group dinner
Close friends: would hang out 1-on-1
Reliants: would ask to help move flats

Out of the group I surveyed, 30% mentioned at least 1 friend, close friend or reliant that I had introduced them to. Certainly makes you feel warm and fuzzy!

Initial survey analysis. My reliants are significantly less likely to have the same social style as I do (5% vs. the survey average of 21%). Also, my reliants are more likely to be first born children (63% vs. the survey average of 51%). Unsurprisingly, there are no ‘tight-knitters’ among my reliants and few among the survey participants.

Simplifications. As you can see from the survey, my friends move a lot. At this point, I don’t acknowledge most of those moves in the visualisation. I also didn’t attempt to visualise when other people met each other or  I introduced them. Finally, I only included a fraction of my acquaintances in order to keep the data set manageable.

The how. There is currently no easy way to do this, but there are a few tools that make it less tedious. Polinode offers an Excel template (1 file with 2 sheets – 1 for nodes and 1 for edges) that will give you a sense of what file you need to create for their software to visualise it.

  1. Start with an existing friends list. In my case, I send quarterly email updates to my personal network, which was a reasonable starting point. Facebook and LinkedIn are likely to be frustrating unless you keep your connections well pruned.
  2. Build a sheet of ‘nodes’ (friends) with at least these basic columns: real name, name, and friend category. I also included email address, gender, nationality, location, how we met, friend categories in time intervals and survey responses. The ‘name’ column is important if you want the real names to remain anonymous. I filled that column in with numbers.
  3. Build a second sheet for ‘edges’ (connections), which is a matrix of who in your network knows each other. Copy the real names and names to this sheet in both the first column and row. You can use a formula to reflect the content from one half of the matrix to the other. Filled out, the matrix it will look something like this: matrix_example
  4. For the node friend categories, I used lost touch, stranger, acquaintance, friend and reliant. I gave them each a value: -1, 0, 1, 2, 3. For the edge categories, I used stranger, known, and introduced. I gave them each a value: 0, 4, 5.
  5. In order to gather the data, I looked at social media and communication tools like email, Skype, WhatsApp and instant messenger programs. Immersion was also really helpful for this.
  6. Once you’ve filled in the sheets, you can save a version of the Excel file without the real names and any other sensitive information.
  7. Next, you can convert the edge matrix to a 3-column table using the ‘reverse pivot’ feature in Excel. There’s a nice step-by-step explanation on Stack Overflow. You can delete the rows that are irrelevant if you like (where the value is either – or 0).
  8. Use Polinode to visualise the network by uploading the anonymised Excel file and selecting ‘No’ to the Advanced Options question about whether the edges are directed (unless you’ve structured your data that way). Once your network has been created, you can play with the parameters to visualise various attributes of the nodes and edges. They have a whole host of tutorials on YouTube.
  9. In this case I created 5 network views, each looking at a particular time slice. Then I exported the pngs and combined them into one image.

Don’t let the short list fool you. Each step takes a lot of effort, but the results are rewarding. Happy to field any questions from people who want to give it a try! As for me, there’s a ton of information to pour over and I’m still learning how to interact with it. Would love to hear any questions you think I should ask the data!

The Reliants Project: 12 months

Can hardly believe it has been over a year since I moved to London! As an anniversary present, London gifted me my first truly serendipitous connection since my arrival. Until then, every new connection was the result of either a direct introduction or meeting at an event that both people intentionally attended. It’s a rare treat to meet anyone during those in between states; by accident, in public places, on transit. I treasure those moments because they often expose a ‘small world’ coincidence or a completely new, fascinating world.

Additionally, I had the chance to participate in Wait But Why’s inaugural Wait But Hi event in August. Our group was even featured in their report (scroll down about 1/5th to “Some people went to restaurants…”). They asked their readers to fill out a (long) survey and then matched them in groups based on their interests and preferences. Some people were set up on individual blind dates while others participated in large group educational seminars (and many variations between). What a fascinating experiment in friendship, relationship and community building!

Round up. Here are some of the more interesting articles and reports related to friendship that have popped up over the last 3 months. First and second round-ups here and here.

It’s official, I have completed a full year of The Reliants Project and have learned an incredible amount along the way – not to mention all the wonderful new friends I have made! Above is a new visual I created with Polinode using my connection-tracking spreadsheet. This only shows connections that were directly introduced to me or have a mutual connection within my personal network. It also doesn’t have any indication of time.

Below I share the final manually-created visual, summarise the statistics, share learnings and propose questions based on what I’ve gathered from actively tracking the growth of my personal network. If you’re looking for my previous posts over the course of the past year+, you can find them here: background, introduction3 months6 months9 months.

 

the-reliants-project-12-months

A quick reminder: the diagram doesn’t keep track of frequency of interactions and I don’t include new connections if there’s no real initiative to maintain contact. I’ve only included co-workers with whom I’ve built a friendship outside of work. I haven’t included anyone I’ve met who does not currently live in London.

Stats. My London personal network has grown about 4x over the last year. That ~90 person network breaks down as follows (again, based on how social networks are often categorised).

  • 50 acquaintances (ie. see them socially)
  • 20 friends (ie. would invite to a group dinner)
  • 10 close friends (ie. often hang out 1-on-1)
  • 7 reliants (ie. would ask to help move flats)

Of the reliants, 4 were pre-existing connections, 2 of which grew from friend to reliant in part due to living in the same city. One of the new reliants was directly introduced to me and I met the other 2 on my own at niche events. Among the close friends, half were directly introduced and the other half I met at events. Friends are disproportionately from my pre-existing network base, but that’s probably because they ‘stood the test of time’ (and distance). I met most connections labeled ‘acquaintances’ at private events, which makes sense because they are part of the same social circle but I haven’t built a meaningful relationship with them individually.

In order to grow that network, I attended about 25 public events (activity-based events, conferences, Meetups, etc.) and about 40 private events (where more than 3 people participated and I didn’t know everyone). I also met with 30 people that were directly introduced to me. Those introductions came from about 65% non-locals contacts, 30% from new local connections and just 5% from pre-existing connections. The graph below shows how many people I met per month over that time.

connectionsgraph_12months

Learnings.

Existing networks don’t directly translate to new connections. Even after a year in London, I have only received a handful of direct introductions from my pre-existing local network. There are still several pre-existing connections that I haven’t even managed to meet with myself! Non-local and new connections were significantly more likely to introduce me to people in their network, as were pre-existing connections that moved to London after me.

‘Super connectors’ increase the number of people you meet, but don’t necessarily result in stronger connections. There are probably 4 people in my network that I would label super connectors. They have each invited me to 3 events where there were at least 15 people in attendance. Only one of the connections I made at these events now falls into my close friend category, but they make up the bulk of my acquaintances.

There seems to be an inverse relationship between size of event and connections made. Conferences, large parties and other groups with more than 15 people seem to make meaningful connections hard to come by. Direct introductions and small gatherings seem much more effective environments for building friendships, particularly if they are hosted by super connectors.

“Weak ties” are strong when building personal networks in new places. If I trace back the 17 reliant and close friend connections I’ve built over the last year, at least 7 have a weak tie at some point in the chain. These connections seem to rely on the social capital of the person making the introduction, where the parties on the receiving end trust that their interests have been kept in mind.

The rhythm of making connections is cyclical. There seems to be a natural balance between building connections and maintaining them. While I focused my time over the first 3 months making new connections, I quickly fell into a pattern of strengthening those connections rather than continuously seeking new ones. Typical holidays, vacations and work schedules also had an impact on the frequency of events where I could meet new people.

Categorising relationships is incredibly difficult and changes over time. While I didn’t actively track the progression of connections from acquaintance to friend to close friend to reliant, a few familiar patterns emerged. The analogy that comes to mind is that of a staircase, where certain catalysts would propel the relationship to a higher step. Most commonly, neither person was motivated to maintain a connection, so we didn’t even become acquaintances. Sometimes there was mutual desire to strengthen the connection, but after several failed attempts to reconnect, the relationship never made it past the acquaintance stage. If the connection managed to avoid those pitfalls, it often developed into a friendship. Significant life events have the potential to be both a positive and negative catalyst, best at bringing people closer together through shared experiences or need for support.

If you’re moving to a new place…

  • If possible, ask your friends (wherever they live) to make direct introductions to people they know in your new place.
  • Attend public Meetups or other events on topics that are of particular interest to you (the more niche, the better).
  • Seek out individuals that are natural networkers and hosts (then attend their events and seek new connections through them).
  • Make time for new connections to strengthen by meeting one-on-one and introducing them to your own network.

What’s next? Right now I’m interested in building a more complex visual of my entire personal network using D3 or Polinode in the hopes that it will give me a more accurate representation of my network than Facebook, LinkedIn or Twitter. I’m also interested in tackling the following questions:

  • How does intention, emotional intelligence and vulnerability impact making connections?
  • Does life stage or the potential for value exchange play a role in which connections are actively nurtured?
  • How do contextual details (e.g. day of week, time, location, structure, event size) impact the effectiveness of making new connections?
  • Can you visualise how new connections transition between acquaintance, friend, close friend and reliant categories over time?
  • Are ‘super connectors’ more likely to have higher measures of betweenness or centrality within their own personal networks?
  • How important is it for relationships within networks to have multiplexity or symmetry to be resilient?
  • How does an individual’s position within their network change as a result of significant life events (e.g. career shifts, relocation, marriage, childbearing, divorce)?

Thanks to everyone for following along!