How to build a new topic

In this article, you will learn how to build a topic using the Community Topics feature

Creating a new topic

Navigate to the Community Topics feature from the Materiality Module menu.

You will land on 'Your Topics' page, which lists all of the topics you have created or leveraged from the Shared Topics library. To create a new topic, type a topic name into the 'Name your topic...' text box and click the 'Build Your Topic' button.

You will land on the 'Topic Details' page. Include a description for your topic in the relevant text box - some additional guidance on formulating this can be found in the template linked below. You are also required to select between 1 and 12 sectors to use as the basis for determining the validity of the topic. The sector(s) that you select should align with the sector(s) you will use in your Materiality Analysis. When all fields are completed, click the 'Save' button to proceed.

Tips for building a high-quality topic

When defining a new topic, it is useful to start by answering a few questions about what the topic represents, how it impacts your company, and where the conceptual boundaries lie. We have produced a template for you to use with some examples to help guide you through this process.

More broadly, there are some rules-of-thumb to bear in mind as you move from the conceptual phase to adding terms:

  • Most emerging or industry-specific topics comprise of between 5 and 40 terms. Terms can be one word (e.g. ‘climate’), a string of 2 or more words (e.g. ‘climate change’ or ‘physical risks of climate change’, etc.), or a string of two or more words located within proximity of each other. The tilde (~) is the symbol used to denote proximity. Proximity can either be specified by the maximum distance between the words or left simply as all words in the term being located within the same sentence.

    To specify proximity, add the tilde followed by the maximum distance between the words – the proximity is applied to all words in the term without preference to their order. For instance, ‘climate risk~3’ could include samples like “climate change presents high risk” or “risk from a shifting climate”. 'Climate finance risks~3' may result in "the finance system incorporates many risks, including climate change." It is advised that the specified proximity be a distance of 5 or fewer words, thereby limiting the potential for false positives - when the term yields a result that is incorrect in the context you expect - in the returned samples.

    For a similar reason, unspecified proximities should be used only in instances when more defined proximity is restricting the results or when 2 or more words in a string must be located consecutively. To use this form, separate words/ strings by the tilde. For example, ‘climate change~risk’ would look for instances in which ‘climate change’ and ‘risk’ are located in the same sentence, regardless of the distance between them - but when the string 'climate change' must be maintained. As with the previous example, this may include “climate change presents high risk” but also “climate change may be important but we have identified labor laws as the biggest risk to our business.” The latter sample could be considered a false positive because the risk is associated to something other than climate change.

    Best practice when considering terms is to start broadly and narrow if there is a high volume of false positives in the resulting samples. Using the previous example of identifying terms to model 'financial risks due to climate change', a broad starting point may be 'financial climate risks~5'. If the results do not accurately reflect what you are looking for, consider making the term more specific, e.g. 'financial risks~climate'.

    When linking a noun to a verb, for instance 'GHG emissions~reduce', consider the different verb formulations that may be used, like 'GHG emissions~reduction', as well as other potential verbs that model the same concept, such as 'GHG emissions~bring down'. Bear in mind the following bullet points as well when undertaking this exercise!

    To add a term, type it into the ‘Add key term...’ text box on the left-hand side of the screen and click the ‘Add’ button.
  • Most topics are driven by a small number of high volume, high frequency terms. These are the ones you want to focus on getting right – they provide the highest return for the least effort – and will be some of the more general, easiest to think of associated with your key concepts. You can assess this using the topic statistics that are displayed after a term is added.

    Screen Shot 2020-02-03 at 12.09.35 AM

    The first set of numbers displayed indicate the total number of reports reviewed as the denominator and the number of those reports in which the term was located as the numerator. In the example, 426 reports in total were reviewed and the term was found in 26 of those. The second number indicates the average frequency of the term within the reports, or how many times, on average, you will find it within each report. In this example, it’s 1.85 references per report. The last set of numbers pertain to your validation, where the numerator is the number of samples you marked as valid and the denominator is the total number of samples reviewed.
  • The process to assess the accuracy of a term in context, or “validate” it, normally requires a review of 15-25 samples, and would not exceed 40 samples. In instances where there is a smaller number of overall samples returned, overall fitness of the term will be determined if at least 80% of the samples are deemed valid.

    After you add a term, you will need to validate it. Click the ‘Validate’ button to open the validation screen. You will thereafter be presented with text extracts - or "samples" - which either accurately include your term in context, or do not. The intention is that you are able to make this determination based on a single sentence in which the term is located, but sometimes additional context may be required. Click the ‘Show Context’ button to display the surrounding text. You can also increase the font size by clicking the magnifying glass icon. To mark a sample as ‘valid’, click the thumbs-up icon; use the thumbs-down icon to mark a sample as ‘invalid. If you are unable to make that determination, you can skip the sample by clicking the relevant button. After you have reviewed enough samples for the system to feel confident in whether the term you selected is going to work for you or not, you will receive a message about the expected term fitness. You can choose to continue validating additional samples, though it’s not expected to change the outcome, and you can always return to re-assess the samples later by clicking the ‘Review’ button.
  • Make sure to remove terms that were not found in any reports or were deemed ‘invalid’ based on your review process. These are identifiable with a red icon and can be removed from the list by clicking the ‘x’ button on the far-right of the term box. If these terms are kept in the topic, they may lead to unexpected results when included in an analysis. You may consider keeping a separate list of all of the terms that were already tried to reference as you progress or at a later point.

After you have established and validated all of the terms in your topic and you are confident in the results, click the ‘Publish’ button on the ‘Term Validation’ page. This will enable you to use the topic in your Materiality Analysis by including it in your Topic Mapping. Find out more here.

After you have published your topic, consider sharing it with the wider community in the form of a template. Just tick the ‘Share this Topic’ checkbox that appears after the topic is published – any updates made by others will not affect your topic.

Once your topic is in ‘Published’ mode, you will be unable to make further edits. If any updates are required, you will need to put to topic back to ‘Draft’ mode by clicking the ‘Return to Draft’ button.