The ability to collect large volumes of data from individuals as they go about their lives is revolutionizing research across a wide range of fields. For the first time, we are able to understand behaviour as it occurs in the real world on a person by person basis. But, conducting lifelogging studies requires new infrastructure and skills. Unforgettable provides a suite of research services to help you through the process.
A comprehensive set of data streams
The number of potential sources of data is growing rapidly. Many current studies focus on one stream (e.g. twitter data), but often a more complete picture can be obtained by triangulating the insights derived from many different sources. Unforgettable offers a comprehensive set of technologies for you to choose from to address your questions.
We have partnered with IFTTT which allows us to collect over 600 additional streams of data. These include calendar information, social media streams, data from wearables and data from internet-of-things devices from brands such as Google, Apple, Microsoft, Fitbit, Facebook, Twitter, BMW, Nike, Pinterest, Instagram, GE and Whirlpool.
Through IFTTT, we can also place buttons on people's smartphones that they push to capture a range of states and events that cannot be passively recorded such as moods, micro experiences (e.g. of aggression or sexism) or behaviours (like taking public transport). We have also interfaced to the SEMA3 ecological momentary assessment system, which issues microsurveys to elicit data from people many times a day.
Participants can download their data from Google, Twitter or Facebook and then upload that data to unforgettable.me for analysis. Our smartphone app can capture accelerometry, audio, GPS coordinates and images on a continual basis and our servers can collect email forwarded from any account.
We also use online databases to enhance the raw data we collect. For instance, we provide weather and temperature information, addresses, place type (e.g. restaurant) and Google Streetview images.
This ain't your grandfather's data analysis. Lifelogging data often accumulates by the gigabyte per month per user or more. Analyzing that data requires specific expertise in signal processing, machine learning and dynamical systems methods. We have created analysis pipelines and training materials to simplify the process. You will find familiar analyses like regression, correlation, and principal components analysis, but also new empirical dynamic methods like convergent cross mapping that allow one to establish causality from time series data. Using our content analysis tools you can analyze textual data like tweets and facebook posts using standard dictionaries, or easily define your own constructs.
When dealing with sensitive data there is a risk that it will be copied and/or mined for personal information (c.f. The Cambridge Analytica/facebook scandal). Business as usual, where data can be freely copied from machine to machine, just isn’t good enough. To address this problem, we have developed a state of the art privacy ecosystem that protects the data from collection to analysis.
Community of participants
Recruiting for lifelogging studies can be hard. The barriers to entry for participants are high and once you recruit them you have to wait for data to be collected before you can start analysis. To address these problems, we have constructed a pool of over 4000 participants, some of whom have been collecting for over five years. In November 2019, we had over 1.3 million events including 416K events from our smartphone app (hour long segments of accelerometry, audio and GPS), 321K emails, 124K SMSs, 53K phone calls and 40K calendar events. But, of course, this is growing every day. You can use our comprehensive costing tool to budget.
And we manage the technology. Troubleshooting the hundreds of phones and services that participants use is the hardest and most time consuming aspect of running a lifelogging study. We have the expertise and manage these issues for you.
Lifelogging data is interesting exactly because it is personal. However, having access to users' personal data raises ethical concerns. At Unforgettable, we take those concerns very seriously and have developed several mechanisms to give users control over their data and make sure it is safe. Our costing tool provides estimates of what participants will be paid, taking the headache out of deciding what to offer and ensuring that participants are paid appropriately for their labour and data. We have navigated ethical review boards in multiple institutions and countries and can help you write your protocols.
It is an exciting time to be a behavioural scientist. We are entering a new phase of development; one which promises a deeper understanding and a new level of translational relevance. Unforgettable provides the infrastructure, so you can focus on the science. To get started, fill in the contact form below or send an email to firstname.lastname@example.org.