Coffee break with CoFounder Aziz
Updated: Jul 22
Today we’ll be shining our spotlight on Capture's CoFounder Aziz, a software engineer who's bootstrapped a software development agency in Islamabad, Pakistan, and Stockholm, Sweden. Ever wondered who made Capture and how?... Read on to hear from Aziz himself!
Capture uses GPS tracking to predict journey time and journey mode so you can automatically track your travel carbon footprint. How does the app know what type of transport you’re taking?
The app uses different data points from the user's device sensors. We look at the activity recognition sensors that are available within most smartphones and many fitness tracker apps rely on them to count a user’s steps. We sample GPS location information only when a user starts rapidly moving to look at their speed and combine it with different services to check if the user might have crossed (or dropped off at) a rail station or a bus stop.
The algorithm behind the app is not yet 100% accurate, but our users can choose to improve the accuracy by providing more data like their typical modes of travel and can also go in and correct a specific entry that may have been incorrectly predicted. What are the calculations behind carbon emissions data within the Capture app? For example for transport, are they localised to each country or region’s transport system’s speed/efficiency/etc.?
All of our emissions calculations for transport are created using official data from the UK Government’s Department for Environment, Food, and Rural Affairs (DEFRA). Once we have mapped a particular trip made by the user to a transport type we use the time spent on that trip to predict emissions. Assumptions around vehicle speeds, load factors (average capacities), etc. are derived from official data from the UK Government’s Department for Transport.
Right now, we are not factoring certain more detailed data such as the locality of a user, energy source for the vehicle (EV, gasoline, etc), or the size of a car... However, this is an area in which we are working to improve.
How much battery energy is consumed when running GPS in the background - did you ever worry about creating an emission tracking app that used more electricity?
We’ve tried to make the algorithm as smart as possible. The location function is mostly kept in sleep mode. We only start looking at the location data when we receive a significant signal of movement. There are intricate differences as to how we handle this across iOS and Android, but preserving the battery has always been of utmost importance to us. Our tests have shown that Capture uses very little battery as compared to most common apps on a phone.
Any data we send to our server is encrypted before being stored. None of the data from our users is shared with any third parties and we have made a commitment as a company to never sell data from our users.
The prediction algorithm runs as a service on our server, where it uses data points from a user’s trip including the location coordinates to predict the type of transport. Once the prediction has been made, this temporary data from the trip is removed permanently from our server.
What’s next? What’s the single most important improvement you would want to make to the Capture algorithm in the near future? (and maybe also the far future?)
The short-term goal is to improve the accuracy of our calculations and give users more control to customise vehicle types. This is one piece we haven’t had the chance to concentrate on yet, but we’ve had some suggestions on using machine learning in the long run to improve prediction accuracy.
What is the next form of tracking you are working on? Could you share a little more on how that will work?
Right now our focus is to tackle sustainability issues at a broader level while working on the Capture for teams module, whereby communities will be able to collectively track savings from planet-friendly actions and commitments. Next on our goal is to expand into tracking emissions from other categories. One of the categories we intend to automate is the home energy usage for users. This would require us to look at energy consumption at a home using the smart meter data and combine it with the emissions produced at the grid level from where the energy is being primarily sourced.
Have any other burning questions regarding how the Capture app works? Do visit our FAQ on our website, or contact us directly at firstname.lastname@example.org!