Habitoclock

What problem does this address?

Knowing how to make your daily routine carbon-aware, which is difficult for the general public.

For the average person, understanding where the energy you use comes from is hard. It's tricky to know what difference you can make to using more renewable sources - and how.

Specifically, it's hard to identify which activities in your day-to-day routine you could change to have the most impact. Lots of people want to know which times of the day are more "carbon dirty" (higher carbon rating on the API).

What is the solution?

An easy-to-use application that helps people understand the impact of their routine. It calculates the impact of their routine activities to provide useful insights, such as what activities they do at different times instead, to reduce their C02 impact.

For example, say you drink ☕ coffee at 13:30, it would suggest saving 11% on Carbon usage by moving it to 10am. We deliver insights like that within a minute - as shown in the demo: https://www.youtube.com/watch?v=4RsoRNPGnOg

It works for all sorts of activities you can describe in text - You can enter any sensible activity. The tool will check whether moving the activity to an earlier or later time will save on the carbon grams consumed. It does that by suggesting moving to times where the carbon rating is lower (= less carbon used for each kilowatt on the current grid). The solution uses an AI third party tool(GPT-3 API) to estimate the watts of activities, so it can work for any normal routine activities - it's not just picking from a pre-existing list, you can add your own custom routine easily! It gives each activity an automatic emoji to make it recognisable, for example: 🍗📱🚿

The Radar graph(shown below) helps understand the difference between actions. It shows what proportion change can be made for each activity. It lists the optimal time for the activity to achieve that with.

The details table(see demo video) shows Watts and the carbon usage at the exact time the user specifies (to the closest 5 minutes on the Carbon Aware API), which is important as high energy usage(Watts) activities should be higher priority.

How does it use the Carbon Aware API?

It requests the rating for energy at the different activity times the user provides. It also checks alternate nearby times for each activity, to find the lowest 'rating' (least Carbon used). 

For example if they make ☕ coffee at 9am, it will check the rating (Carbon grams per KW hour) at 9am, and times around 9am, such as 8am, 11am, 1pm, etc. See the files HabitStore.js and CarbonUsage.js  for this.

Our process uses a backend API to gather the estimated watts used by the activity for 10 minutes, using OpenAI' GPT-3, which we find is accurate for most activities (and rates 0 watts for non-activities). There are some edge cases, but it is impressive. To help users distinguish between activates, we use GPT-3 to assign an emoji to the activities.

For example, after all the processing, we go from just the activity and times of "charge 📱 Phone at 10:15" and "make ☕ coffee at 13:30" to working out the watts, carbon usage at that time, and suggesting better times (in this case, charge phone earlier and drink coffee much earlier). It will also recommend time-shifting activities to later in the night often (assuming the dirty energy peak is in the afternoon, as typical in uksouth).

Describe the impact?

It would have high reach, and low intensity. Primarily, it would educate and enable people in the general public to know and act upon their C02 impact. It would help people immediately lower their C02 impact.  Let's say, practically, if 1 in 5 people used the app, and each followed 1 in 3 suggestions, it would reduce overall activity-based C02 usage by 1-2%.

More than this, there are psychological benefits to this approach to routine change that is not seen with say a table listing watts and times of dirtier Carbon rating values. You're more motivated to change habits for the better when it helps the world, not just yourself - and when you feel you are learning useful sustainability information about energy usage of various activities and when energy is dirty.

🤸 1 in 5  people using 1 in 3 suggestions

🧍🧍🧍🧍🧍 = 1-2% C02 reduction possible 

Would people choose to use it?

I think a lot of people would, especially in 2 months on new years! I think the awareness potential is the strongest virtue, because this combines estimates on how much each activity uses in terms of carbon with the timing data - learning that information is valuable in itself as even as you fill in the form you get useful information at watts used by activities: I was interested myself and my friends to use it and understand energy, it is valuable for that reason alone.

🎇 New Years Resolutions is perfect timing

Is it feasible in reality, and to scale?

It is highly feasible, as it's scalable with the Carbon Aware API and a quick simple frontend. It would be used at equally distributed times so no major overload stress issues. The 1st of January is a perfect kickoff time to expand widely as lots of people will be thinking about routines and new years resolutions, which directly provides motivation for. Earlier on in October, we tried a chat UI, but it actually felt somewhat dystopian, so we switched to the form input. So we think this UI works well. We find it is fairly robust to different user inputs but would like to connect to some official watt databases and localised information to improve the accuracy of the tools energy estimation.

It's easy for others to run our code once you specify the secret Open AI API key and ensure the Carbon Aware API is working from our github code.

What vision do you have?

The vision is to make this available to the wider population so that they can understand their impact and make time-conscious choices about energy usage. In 2016, I ran a "Green Design Contest" at the University of Sussex (Informatics, Design and Engineering Society) and the top submission was a site that helped students learn the map of the campus university while learning what you can recycle in each bin.

It was a brilliant design, but why?

Because students were naturally motivated to learn the map to get to know their new campus - and without effort, just during that, learnt to recycle correctly. That change bettered the environment (recycling more correctly).

Imagine the vision of applying this product here on New Years Eve in a more gamified way, helping people learn their energy and carbon impact while setting resolutions (e.g. give up TV, drink less coffee). It'd be a perfect way to educate and enable people to consider Carbon impact, while they are highly motivated by their own resolutions!

The timing with 1st of January is perfect and as it is scalable it could be deployed easily to thousands of users with a competent set up within a month.

In the future progress tracking could allow you to motivate as you build up a "C02 streak", just like Duolingo or Snapchat, but with real world positive impact. Most importantly of all, we can offer this to the population at large as it's easy to use. It could definitely have the same look, feel and vibe or the Forest app (which plants trees as you use it positively, hooking users).

Where is the code/prototype?

Github code is here: sussex-james/habitoclock (github.com) 

We don't have a prototype live because you need a sensitive secret API key but we can host a live demo as needed happily and share screen!

Footnote

Thank you for reading this application!

The makers of this project are James and George who have released the code on Github:  sussex-james/habitoclock (github.com) 

Feel free to ask any questions to us.