We're changing how we measure your environmental impact 🚨

If you’ve ever created an event or entered repair data on Restarters, you’ll know that one of the most powerful features of the Restarters platform is estimating the environmental benefits of each repair event and community repair group.

We’re working on an exciting update that will improve these estimates. These are big changes that will affect how your environmental impact is calculated. So, this will be quite a long post outlining exactly what’s changing.

Here’s the short version: we’re using new environmental data we’ve sourced and adding support for more types of item to greatly improve the accuracy of the environmental stats generated for your group and events :straight_ruler:. And for the first time, we will also add support for unpowered items, such as clothes :jeans:.

It’s important to know that because of the way the software works, these changes will also affect your existing impact stats (if you have added repair data to any of your events). But for nearly 90% of groups with repair data, your overall CO2e savings are actually set to increase :chart_with_upwards_trend:.

We plan to release this update in October. But in the meantime, keep reading for a more detailed overview of what’s changing and why…

click here for a short summary of the posts below
  1. New environmental data - we now have much better data about the carbon footprint of common household products. This will improve the accuracy of our calculations and let us add estimates for unpowered items.
  2. Impact stats for miscellaneous items - the way we currently calculate these means your impact stats may change over time, even if you don’t add or remove data. This update will prevent this kind of ‘drift’.
  3. New product categories & recategorisation - we’re adding environmental calculations for 10 product categories, 7 of which are completely new (e.g. sewing machine). We’re also planning to reclassify items that are currently ‘miscellaneous’ into the new categories where appropriate.
7 Likes

1. New environmental data :bar_chart:

To estimate the environmental benefits of repairs, we rely on reference data about the CO2e footprint of common electrical and electronic devices.

:laptop: For example
We know that an average medium-sized laptop produces around 260 kg of CO2e before it’s ever used. If you help someone repair one of these laptops at an event, we assume that you’ve extended its life by about 50% - and by doing so, you’ve saved the owner from buying a new replacement for that much longer. This equates to about 50% of the total impact of producing a new laptop: around 130 kg of CO2e.

But the data we’ve been using is now getting old. It was first gathered in 2014-15 and products have changed since then. We were also missing high-quality data for quite a few types of product.

By now, you’ve probably heard that we’ve recently completed a big project to update our environmental data. If not, you can read all about it here. In any case, here’s the important bit: we now have much more data on many more types of products than before. This will allow us to make more accurate, product-specific CO2e estimations for most items you see at repair events.

But what about products that we don’t have specific data for…?

:arrow_down_small: continues below

4 Likes

2. Impact stats for miscellaneous items :pushpin:

People sometimes bring along rarer or more unusual items to be fixed at events. And even our new environmental data doesn’t cover every single type of product that exists. So when you add one of these ‘miscellaneous’ (or ‘misc’) items to your event in Restarters, we come up with a rough estimate for a generic powered or unpowered item. Here’s how it works at the moment…

Current method
Together, we’ve all recorded a whopping 20,000 repair attempts at community events! This gives us a pretty good idea of the kinds of items we all see most often at events. To work out the rough impact of repairing something that doesn’t fit one of our product categories, we use data about all the products that do fit to produce a CO2e to weight ratio:

The rough CO2e impact per kg of weight for misc powered devices =
Total CO2e of all (non-misc) powered devices in the Fixometer /
Total weight of all (non-misc) powered devices in the Fixometer

This is why we ask you to record the weight of misc items. Once we have the weight, we apply this CO2e to weight ratio to estimate the item’s carbon impact. While not perfect, this method is usually enough to produce a reasonably fair estimate.

But there’s an issue. Because it relies on calculations on the whole database, this ratio changes over time. If someone adds a few toasters to their event, the ratio will drop slightly for everyone. This means that if you have added misc items to any of your events, your overall impact stats can change even if you haven’t added or removed any data yourself.

To prevent this, we’re changing how we make this calculation.

New method
We’re planning to introduce a fixed CO2e to weight ratio. We can do this now because we have so many items in the Fixometer and better environmental data. Over time, the ratio we used has become more settled, allowing us to work out a representative ratio that we can be confident in.

Using a fixed value means that your impact stats will no longer change based on the repair data from other groups; after the update, only changes you make to your data will affect your stats.

:arrow_down_small: continues below

4 Likes

3. New product categories & recategorisation :twisted_rightwards_arrows:

Finally, we’re also adding new impact calculations for 10 product categories, 7 of which are completely new to Restarters. They are:

New Powered Categories New Unpowered categories
Coffee maker :coffee: Bicycle* :bike:
Games console :video_game: Clothing/textile* :jeans:
Iron :dash: :tshirt: Furniture* :chair:
Sewing machine :thread: Hand tool :hammer:
Watch/clock :watch: Jewellery :ring:

* these 3 categories already exist, but we are adding CO2e calculations for the first time

These new categories should help you more accurately record the items seen at your events in future and you’ll get more accurate environmental impact estimates.

We chose these based on how often these types of items are brought to events according to the data we’ve all recorded in the Fixometer (remember MiscCat, where we looked through ‘misc’ items and re-categorised them?).

That obviously means there are quite a few of these items already in the Fixometer, currently in the ‘misc’ category. So, as part of this update, we will also be putting these items into their new, correct categories; relabelling every coffee maker from ‘misc’ to ‘coffee maker’, for example. As a result, any of your items that are recategorised will be given new environmental impact estimates, which will affect your overall stats.

Do you have questions? Keep reading…

:arrow_down_small: continues below

4 Likes

FAQs

How will my group’s stats be affected?

These changes will not only affect how your environmental impact is estimated in the future, they may change your current stats. This will be slightly different for every group:

  1. Powered items: the new environmental calculations will apply to new and existing data. For most powered categories, the new impact measurements are higher than in the previous version of our source data. So, for most groups, the recalculated impact stats will come out higher too.
  2. Unpowered items: we are adding impact estimates for these items for the first time. So, if you have recorded unpowered items, your overall impact will increase.
  3. Miscellaneous powered items: the stats for your miscellaneous powered items (anything without a product category) will not change. In fact, they will stop changing.
  4. Recategorised items: some of your miscellaneous items may be recategorised into one of the 7 new categories. For these items, the impact may increase or decrease depending on the new category. For example, watches will decrease slightly, but games consoles will increase significantly.

Overall, our modelling shows that CO2e savings will increase for around 87% of groups and waste savings will increase for around 55% of groups.

:arrow_upper_right: For groups whose CO2e savings increase, the average increase will be around 25%.
:arrow_lower_right: For groups whose CO2e savings decrease, the average decrease will be around 15%.

:arrow_upper_right: For groups whose waste savings increase, the average increase will be around 27%.
:arrow_lower_right: For groups whose waste savings decrease, the average decrease will be around 10%.

If you would like a specific estimate for your group, just post below :slight_smile:

Why are you planning to apply the new calculations to existing repair data?

The new data we’ve found gives us a more realistic idea of the embodied carbon of the kinds of products we see at events, even for events that happened years ago. For example, for large laptops, we only had one reference example. Now we have 16. And for some types of products, we can now use category-specific reference data for the first time, such as some sizes of flat-screen displays. We’d like to bring this new detail to the data already in the system.

On a technical level, we’re also finding that it would be much more simple to update all calculations, rather than build and maintain two different systems. Using the old calculations for existing data would also raise a number of complex issues, such as what to do with newly-entered data for past events, how or whether to use the new calculations for existing data in the new categories and so on. Ultimately, we’d like to keep the system easy to maintain and easy for everyone to understand.

More questions?

Just post below

4 Likes

We’ve just fixed our first electric bike, which I think will give us a categorisation problem…

Bikes are an interesting item in that people bringing in bikes aren’t usually looking for a welding job, it’s normally a component or accessory repair, e.g. tire or lights and they are unlikely to throw the whole bike away if the lights can’t be fixed.

Some example categories…

Bike light -> powered -> Decorative or safety lights
Bike battery -> powered -> Misc -> estimate=3 kg (or whatever) -> Item type=“eBike battery”
Bike tire -> unpowered -> Bicycle -> estimate=0.5 kg (or whatever) -> Item type=“tire”

Note that for

  • powered and unpowered items categorised as Misc, calculations require an estimate
  • powered non-Misc items, calculations will ignore any estimate entered, the weight of the category is used always
  • unpowered non-Misc items, calculations will look for an estimate first and fall back to the weight of the category if there is no estimate

Does that help?

1 Like

We’ve had two flat screen TVs in this month…both 40" but there isn’t a category for that…have put under 32 - 37, but wanted to mention. People seem to require ever bigger tvs these days.

2 Likes

Good point @Clare_Seek. In past we extrapolated and interpolated the impacts of screens we didn’t have reference data for, based on available data. I’m sure @james can answer this question when he’s back from some annual leave.

This is great and looking forward to it. Can we also look at “small kitchen item” which is misc category for blenders, mixers etc.

Um… Dunno. We had a fault in the electric drive, where it failed to run properly. Bike was still useable, but only used electric power intermittently. We stuck some sensors back in place. Weight minimal, so is that what our estimate should be?

If the electric drive had not been repairable would it mean replacing a larger part? If so, then the weight of the larger part would be the estimate, otherwise the weight of the electric drive itself would be the estimate.

For the Open Repair Data Standard we have introduced “Coffee maker” and “Food processor” - the latter includes all small kitchen items that consist of a motor that rotates a blade to chop or blend something. Introducing these into the Fixometer is under consideration, for now please use the item_type field to record “Blender” etc.

When adding new categories to the Fixometer there are a few things to be considered, such as the length of the pick-list of categories which will get unwieldy, especially on smaller devices, and the number of choices can cause confusion resulting in wrong selections and people being put off entering their data when it becomes too much of a chore.

The Fixometer devs are currently working on a method of auto-suggesting appropriate categories based on item_type and other factors in the hope that this will mitigate some of these issues. For now, item_type is the best place to describe the actual item and when we do introduce new categories we use that field to re-assign the category.

2 Likes

Seems good/logical, but perhaps we need some special guidance re. bikes, based on what you’ve shared? I think @Robert_Irving has a point, it could confuse a lot of people, especially non-technical people entering data.

1 Like

Indeed, there needs to be an FAQ covering how-to categorise and how/when to estimate. It should be accessible - and prominent - from within the data entry form.

1 Like

Hi everyone - a quick update. We’ve been continuing work on this update and hope to release these changes in the next week or two. We’re excited for you to see what we’ve been working on and start seeing improved impact stats for your events!

I’ll post back once this all goes live, so keep an eye out for further news :slight_smile:

2 Likes

Another update from me.
We will be updating your environmental stats 2021-10-27T09:00:00Z!

:warning: Please note that Restarters will be offline for about an hour while we complete this work

During that time, you can reach us at: community@therestartproject.org :slight_smile:

We’ve now completed this update! :tada:

You can now see the updated stats for your group and events!

And for a breakdown of how environmental calculations work, take a look at our new help topic:


:question: Are you seeing lots of ‘NaN’ where the stats should be? This might mean you’re seeing an older version of the page. Try force-refreshing the page or clearing your browser’s cache.

1 Like