Carbon GPT
User Guide

Step 5: Collect activity data

Written by Firdaus Adib (AI Product Manager) Reviewed by Dr. Ivan Ling, PhD (Chief Scientific Officer & Co-Founder)
Published Updated

This is step 5 of 7 in Build your GHG inventory. Previous: Step 4: Identify your emission sources. Next: Step 6: Calculate your emissions.

Activity data is the measured amount of each activity that causes emissions, such as litres of diesel or kWh of electricity. In this step you collect it for every source on your list, for the whole reporting period. Good activity data matters more than anything else for an accurate inventory.

What data each source needs

SourceActivity dataUsual unitWhere to find it
Boilers and generatorsFuel usedlitres, m³, kgFuel invoices, delivery notes, tank readings
Company vehiclesFuel used, or distancelitres, kmFuel cards, fleet logs
Refrigerant leaksRefrigerant added during servicingkg, by typeMaintenance records, service invoices
Purchased electricityElectricity usedkWhElectricity bills, meter readings
Chilled water or steamEnergy usedkWh, GJ, RT-hSupplier invoices
Purchased goodsQuantity or spendtonnes, RMProcurement system, invoices
FreightWeight and distancetonne-kmLogistics records, freight invoices
WasteWeight by type and treatmenttonnesWaste contractor reports
Business travelDistance by mode and classpassenger-kmTravel agency reports, expense claims
Employee commutingDistance, mode and dayskmStaff survey

Fuel quantity is better than fuel cost, because fuel prices change. For refrigerants, the amount added during servicing is a common way to estimate how much leaked.

Primary and secondary data

  • Primary data comes from your own operations or from your value chain partners: meter readings, invoices, supplier emissions figures.
  • Secondary data comes from averages: industry data, databases, or spend-based estimates.

Use primary data for Scope 1 and Scope 2. For Scope 3, primary data is best for your largest categories. Secondary data is acceptable for smaller ones and to fill gaps. IFRS S2 asks you to explain how much of your Scope 3 uses data from specific activities in your value chain.

Judge the quality of your data

The Scope 3 Technical Guidance uses 5 tests of data quality. They also work for Scope 1 and Scope 2:

TestQuestion
TechnologyDoes the data match the technology that you actually use?
TimeDoes it come from the reporting period?
GeographyDoes it match the place where the activity happens?
CompletenessDoes it cover all sites and the whole period?
ReliabilityIs it measured or verified, or only estimated?

Methods for purchased goods and services

For Scope 3 category 1, the Technical Guidance describes 4 methods, from the most to the least specific:

  1. Supplier-specific: product emissions data from your suppliers
  2. Hybrid: some supplier data, with averages for the rest
  3. Average-data: the weight or quantity of goods multiplied by an average factor
  4. Spend-based: the money spent multiplied by an average factor per RM or USD

Many companies start with the spend-based method and move to better methods for their largest suppliers.

Plan the collection

Turn your source list from step 4 into a plan. For each source, record who provides the data, in what format, how often, and by when. Monthly collection helps you find errors early. Annual collection is easier but leaves less time to fix problems.

When data is missing

  1. Find the gap. Note which site, source and months are missing, and why.
  2. Judge its size. Estimate how much it could change the total.
  3. Fill it. Use data from other months, a similar site, or an average, for example kWh per square metre of office space.
  4. Record it. Write down the method, so that a reviewer can follow it. Replace the estimate with real data when you can.

Keep the evidence

Keep each invoice, meter reading and survey result, and note where the number came from. An assurance provider will ask for it. Good practice in carbon accounting explains what an audit trail needs.

Check your progress

Before you go to step 6, make sure that you have:

  • Activity data for every source, for the whole reporting period
  • A note of the data source and quality for each figure
  • A record of each gap and how you filled it

In Carbon GPT

Record each item of activity data as a measure in Measures, with its date, facility, amount and unit. To collect data from colleagues or suppliers, send a form with Requests. Their answers come back as measures after you approve them.

Sources