• The data for the CFT comes primarily from Crunchbase.com and Candid.org, with supplemental details from LinkedIn and company websites.

  • We compiled a list of over 279 terms related to climate mitigation and climate adaptation based on climate finance reports (e.g. Climate Policy Initiative - state of global climate finance), climate synthesis websites (e.g. Project Drawdown, ReGeneration), and Wikipedia pages that mention ‘climate mitigation’, ‘climate adaptation,’ or ‘climate resilience’. We searched Crunchbase and Candid for any US-based investee or grantee that mentioned any one of these terms and received at least $50,000 total funding in the past 5 years. These searches cast a wide net and included many false hits - for example, an organization that mentioned ‘climate’ but it was used in the wrong context (e.g. ‘in this political climate…’). To filter out false hits, we labeled data and fine-tuned a large language model from OpenAI (babbage GTP-3) to detect if an organization’s work was actually climate-relevant or not. The fine-tuned model achieved between 90-95% precision and 92-94% recall on the validation sample. For more details, see Methods.

  • Every data point on the Tracker represents an organization, either a company or nonprofit whose work relates to climate mitigation or adaptation. The clusters represent groups of organizations whose descriptions share similar climate-related keywords and, together, self-define an emergent theme. The groups and themes are not defined in advance, but rather coalesce automatically - using machine learning - from the shared language of how each organization describes what they do.

  • Clusters are formed nondeterministically - they coalesce based on the language used across the entire dataset. The presence of one keyword in an organization’s description doesn’t dictate which cluster the organization will end up in. All the keyword terms within an organization’s description are considered and compared against the usage of those keyword terms within the entire dataset so that an associated weight can be determined for each keyword. The similarity of each organization’s descriptions is compared, and clusters are formed based on keyword similarity, taking into account the weight of the keywords present within a description.

    • This resource is still a work-in-progress! We are relying on various technologies to automatically identify relevant organizations, and we know these methods are not all-encompassing or perfect. Irrelevant organizations sometimes get past our filters, and relevant organizations are sometimes missed.

    • Also, this resource is not intended to be a comprehensive catalog. It is not a ‘census’ but rather our best effort at mapping a representative sample. As is, it is intended to reveal overall trends rather than serve as a discrete database of all climate-related organizations.

    • That said, we are constantly improving our methods and would love your feedback. If you notice something missing or something that doesn’t belong, please let us know!

    • We are relying on various technologies to assemble the dataset and are constantly improving our methods. If something is wrong, please tell us!

    • However, in some cases, the data is only as good as the underlying data sources. Check if the underlying data is wrong in Crunchbase or Candid or LinkedIn. If so, they’d love to know! Update your organization’s profile on their sites or contact their support teams for help.

  • There is not currently a process for users to make any data updates or data additions on the fly, but please let us know if you think something needs to be deleted, changed, or added! Have a whole separate data stream to add? Let us know!

FAQ

DATA SOURCES, ISSUES, & INFORMATION

HOW TO USE

  • Watch this video for a short tour and tutorial. The CFT is an exploration and discovery tool. The dataset is a representative but inexhaustive compendium of companies and nonprofits doing work in the climate mitigation or adaptation space and includes a rich set of associated metadata for each organization. The Data Filters on the left (under the Summary tab) enable you to slice and dice the data any way you want. The Snapshot Selector at the top offers a selection of saved presets that enable you to explore different perspectives of the same data. You can click on an individual data point to view organization details. You can search via the Search bar at the top. Learn more about each feature in depth below.

  • The Snapshot Selector offers a selection of saved presets for exploration.

    • Climate Keyword Themes: Organizations automatically clustered (using machine learning) into themes based on keyword similarity.

    • Climate Equity vs Venture: Climate Keyword Themes sorted by the fraction of orgs within that mention equity terms vs the fraction that are venture funded. Data points are colored relative to the fraction of organizations within the Keyword Theme that are equity-related. The darker the color, the higher percentage of equity-related organizations within the Keyword Theme cluster.

    • Geographic View: Organizations plotted geographically based on office locations, colored by Climate Keyword Theme.

  • The Legend tab displays what attribute the data points are clustered by, colored by, and sized by in the data visualization.

    The Summary tab displays the data attributes associated with the data currently selected for analysis.

    The List tab displays a list of all data points currently selected for analysis.

  • Clicking one of the elements in the list of Data Filters selects all organizations that have that data attribute. The selection will reflect on the data visualization but won’t change the options in the Data Filters list. The options will change if you click ‘Summarize Selection’ to reflect the data elements associated with your selected set.

    Selecting multiple values within a category acts as an OR operator (for example, if you select ‘energy’ and ‘water’ within the Keywords category and then click ‘Summarize Selection’, the results will contain any organization associated with the keywords energy OR water).

    Selecting multiple values across categories acts as an AND operator (for example, if you select ‘energy’ in the Keywords category and ‘low income’ in the Equity Tags category, the results will contain any organization associated with the keyword ‘energy’ AND has the ‘low income’ equity tag).

  • Clicking ‘Summarize Selection’ after you’ve selected one or more Data Filter(s) results in a reset of the Data Filters options. The list of Data Filters resets to display the underlying data characterics of your current selection. “Summary” of the selected data.

  • Searching via the top search bar performs a text search on the data attributes. If you only want to search on specific fields, you can use the Search Filter.

  • Clicking on one data point in the graph displays the details for that particular organization in the Information panel on the right. Clicking whitespace will clear your selection.

  • Top-level information about the CFT can be found by clicking the info icon in the top right corner and opening the Information panel. If a data point is selected, organization data will be displayed, otherwise, top-level info will display.

  • Unfortunately no, there is not a way to export data from the CFT. All the code to generate the tracker is open source, but the data in this map are owned by Crunchbase.com and Candid.org. They generously agreed to provide the data and make everything searchable and accessible in the map - but we do not have permission to allow exporting of the raw data.

HOW TO GIVE FEEDBACK OR FIX/ADD SOMETHING

  • Fill out this Feedback Form or contact us at support@vibrantdatalabs.org. The CFT is a living tool that we’re actively working on, and we need your help to make it better!

  • Is there something wrong with the profile details for an organization? You can contact Crunchbase or Candid directly if you see that an organization’s metadata is wrong on those sites. Otherwise, let us know! Fill out this Feedback Form or contact us at support@vibrantdatalabs.org.

  • Fill out this Feedback Form or contact us at support@vibrantdatalabs.org if you have an organization, relevant metadata, or a data stream that we should add to the CFT. There’s currently not a way to add data on the fly, but if there’s something we should include, we want to know about it! And if you think we should build a way for you to add data on the fly, we also want to know that!

  • We’d love your input! Fill out this Feedback Form or contact us at support@vibrantdatalabs.org. Would you be willing to participate in a small focus group? Let us know!