Pick the kind of sustainability work you want to demonstrate
A sustainability internship can involve energy data, reporting, supplier questionnaires, product research or communications. These tasks do not require the same portfolio. For carbon-accounting and reporting work, a useful sample is a traceable activity-data workbook and a clear methods note. A visually appealing sustainability presentation is less persuasive when its figures cannot be followed back to a defined boundary and source.
The GHG Protocol Corporate Standard provides a framework for organizational emissions inventories. Its overview distinguishes that purpose from accounting for project reductions used as offsets or credits. The Scope 2 Guidance addresses purchased electricity, steam, heat and cooling. These are accounting references for learning how to frame the work, not a declaration that any calculation bearing their names is compliant.
Search sustainability reporting intern, carbon accounting intern, environmental data intern and corporate sustainability intern. Read the assignment to see whether it centers on inventory data, disclosures, operational projects or communications. A role requiring engineering work at a facility should not be treated as a spreadsheet-only assignment. A communications role should not be represented as technical emissions assurance.
Define the organization and period before collecting numbers
For this fictional project, define one organization operating one office for a full twelve-month reporting period. State which activity records you expect to receive and who would own them: purchased electricity invoices, a fuel-use record and documentation of any boundary assumptions. The narrow scope makes it possible to reconcile the example without pretending to cover a multinational business.
Keep a boundary register. Which facility is included, what dates does the organization control or occupy it, and what is the basis for attributing shared energy use? A bill addressed to the building owner is not automatically entirely attributable to one tenant. A new office added during the year needs an explicit treatment rather than being inserted into a comparable-year chart without explanation.
Write unknowns as questions with consequences. If a shared meter cannot be allocated reliably, identify the missing information and avoid presenting the allocated figure as a direct measurement. If a record covers only ten months, do not label its total annual consumption. A transparent incomplete inventory is more useful for a practice project than a complete-looking number made from hidden assumptions.
Build a source ledger that can survive review
Use one row per source record with a facility ID, meter or account ID, period start and end, activity quantity, unit and document reference. Add fields for actual versus estimated data and the reason for any correction. Keep the original values beside cleaned values. That makes it possible to investigate a suspicious result without trying to reconstruct what changed.
Consider common errors in the fictional office data. An invoice might include a previous-period adjustment; two files may contain the same meter period; one record may report MWh while the rest use kWh. These are different problems. Deduplicating on the amount alone could delete two valid equal bills, while summing mixed units without conversion could distort the result by a factor of a thousand.
Include simple reconciliations: do the covered periods span the intended year, are there overlaps, do known facility totals agree with the supporting records, and are missing quantities distinct from genuine zeros? A blank cell means information is missing unless your documented process establishes otherwise. Do not turn every blank into zero merely to make a total calculate.
Work the units through a small fictional inventory
Assume annual electricity consumption of 120,000 kWh and a purely fictional factor of 0.35 kilograms of CO2-equivalent per kWh. The result is 42,000 kilograms, or 42 metric tonnes of CO2-equivalent. Separately, assume 10,000 kWh of fuel energy and a fictional factor of 0.20 kilograms per kWh, giving 2 metric tonnes. The two modeled components total 44 tonnes under this exercise's stated boundaries.
The factors are teaching inputs, not recommended factors for a real location, year, fuel or reporting method. In a real assignment, record the factor's source, version, geography, applicable period and units. A factor expressed per unit of fuel volume cannot be multiplied by energy without the appropriate conversion and its assumptions. Avoid silently substituting whichever factor makes the result smaller.
Keep each component visible. The electricity calculation does not cover purchased goods, commuting or every other possible activity. The fuel example does not establish the classification of a real leased asset. A 44-tonne sum of two defined components should not be labeled the full footprint of a company whose remaining activities have not been investigated.
Separate operational changes from changes in the factor
Now assume electricity consumption falls 10% to 108,000 kWh while the fictional factor remains 0.35. Electricity emissions in the model become 37.8 tonnes, a reduction of 4.2 tonnes. If the fuel component is unchanged at 2 tonnes, the modeled combined total becomes 39.8 tonnes. That is a reduction of about 9.55% across these two components, not 10% across a complete corporate inventory.
A different scenario changes only the electricity factor to 0.30 while consumption stays at 120,000 kWh. The electricity result becomes 36 tonnes, but the organization did not use less electricity in that scenario. A reporting note should explain whether the result changed because of activity, a factor, a boundary change or a data correction. The same headline percentage can hide very different explanations.
GHG Protocol's Scope 2 material addresses contractual instruments and their quality criteria for the market-based method. Do not replace electricity consumption with zero merely because a document mentions renewable energy or an offset. Keep the relevant methods and evidence distinct and use current guidance and qualified review for a real inventory. This exercise does not determine the treatment of an actual energy contract.
Sources[2]
Write a methods note as carefully as the chart
Your methods note should identify the reporting boundary, period, activity sources, conversions, factors, exclusions and unresolved records. Explain which figures are measured and which are estimated. Add a change log so someone can distinguish a correction to the dataset from an operational reduction. A favorable result should not be easier to find than the definition that limits it.
In the fictional office case, a useful conclusion says that modeled emissions from the two included energy components fall from 44 to 39.8 tonnes under the reduced-electricity assumption and fixed factors. It does not claim a measured reduction, climate neutrality or a verified company-wide footprint. The statement is still useful because it tells a reader exactly what was calculated.
A good sensitivity table can show how uncertain activity or factor inputs affect the result. Keep this separate from confidence intervals unless you have a defensible statistical model for those intervals. A low/base/high scenario is a set of assumptions, not automatically a quantified probability range. Naming the uncertainty correctly is part of the analytical work.
Package the project for a reporting or data internship
Submit the synthetic source ledger, a calculation sheet, a missing-data log and a one-page methods note. Add a small chart only after the totals reconcile. The hiring signal is that another person can inspect your work, reproduce the result and identify what should not be inferred from it. An elaborate dashboard is optional; traceability is not.
In an interview, explain an error case such as mixed kWh and MWh, duplicate periods or a change in the organizational boundary. Describe how the issue would be investigated and how the correction would be documented. Use fictional or public-source data rather than sharing a previous employer's bills, supplier information or private reporting files.
Ask whether the internship's main deliverable is a cleaned dataset, inventory support, disclosure research or an operational analysis. Clarify which standard and review process the team uses. Your preparation should make it easier to contribute to that specific deliverable, not imply that a short independent project qualifies you to certify a company's emissions reporting.