Skip to main content
Getting data recommendations

How to get the best data recommendations from Sidekick

Jennifer Funk avatar
Written by Jennifer Funk
Updated over a week ago

One of the key features of Sidekick is its ability to analyze your natural language query and recommend the most relevant datasets from the mySidewalk data library. Simply describe your topic or question, and Sidekick will suggest specific datasets that can help provide insights.

When to use this capability

When you need data to analyze for a particular project, question, or analysis you're doing, it can quickly point you to the best datasets to explore — saving time spent having to search through thousands of datasets on your own.

Example use cases include:

  • Writing a grant proposal that requires data on the issue being addressed

  • Conducting a needs assessment to identify priorities for programs/services

  • Developing a strategic plan that incorporates data on organizational focus areas

  • Evaluating the impact and outcomes of an existing program or initiative

  • Preparing a policy brief or report to inform decision-makers on a local issue

  • Exploring data for a presentation or proposal to stakeholders/funders

Example questions and responses

User Query

What data should I look at to explore the racial wealth gap in my community?

User Query

I'm working on a grant proposal to get funds for affordable housing development in our city. What data should we include?

User Query

What data should I look at to understand mental health conditions and access to care in my community?

Advice for effective prompting

To get the most relevant dataset recommendations from Sidekick, follow these tips when framing your query:

  • Clearly state your project focus. Good context for Sidekick includes the topic and goals of your project, as well as the target population, if applicable. For example, "I'm working on improving healthcare access for individuals at risk of diabetes. Recommend datasets that can guide my efforts."

  • Specify the geographic area if relevant (city, county, ZIP codes, etc.). By doing this, you can help ensure that Sidekick recommends data that is available not only for the place you’re interested in, but also at the level of granularity you desire.

  • Don't make several requests in the same query. Sidekick works best when asked clear and discrete questions. When asked to perform several requests within the same query, it can return mixed results. Instead, try asking one request per query, and follow up with additional questions as needed.

  • Use additional features for deeper insights. As a follow-up to basic data recommendations, consider asking Sidekick to answer a question or generate a map from the recommended datasets. This can enhance your understanding of the data points and help in presenting them effectively in reports or presentations.

Did this answer your question?