Engaging the Scientific Community in the Use of Gridded Population Data

Written by Rebecca Gorin, SDSN TReNDS’ Intern

As the world continues to face the immense challenges posed by Covid-19, nontraditional data sources, including gridded population data, play an important role in tracking the virus’ spread and ensuring that no one is left behind. Gridded population maps distribute data using grid cells, combining census results with additional information, such as geospatial data from satellites, to provide more accurate and timely population estimates. However, as our recent report on gridded population data and joint webinar with the International Science Council revealed, while they offer great promise, many policymakers and researchers are still largely unaware of gridded population data, particularly their nuances and potential applications for sustainable development and Covid-19.

The webinar, “Accounting for Everyone: Using Gridded Population Data for Sustainable Development,” highlighted the experiences of policymakers and researchers working with gridded population data and was moderated by TReNDS’ Director, Jessica Espey, and featured TReNDS’ expert member, Lisa Bersales, and analyst, Hayden Dahmn, along with representatives from the Center for International Earth Science Information Network (CIESIN), and the Global Partnership for Sustainable Development Data (GPSDD).

Five key takeaways emerged from the discussion:

1. There is significant potential for gridded population data in the fight against Covid-19.

When a disaster strikes and resources need to be mobilized rapidly, gridded population data can help policymakers better understand where vulnerable populations are impacted in a short period of time. For example, the World Food Programme (WFP) is able to respond more rapidly to emergencies using LandScan and WorldPop datasets to estimate near real-time impacts of earthquakes and tropical storms. In the context of Covid-19, CIESIN’s Susana Adamo highlighted the NASA Socioeconomic Data and Applications Center’s new interactive Global COVID-19 Viewer, which uses gridded population data to disaggregate virus cases by age group and sex to better understand the spread and severity of the virus.

According to GPSDD’s Victor Ohuruogu, in Africa, Covid-19 has generated more demand for gridded population data. He explained that prior to the current crisis, many statistical offices had not previously engaged with gridded population data, but now there is growing familiarity and use. For example, some countries, like Nigeria are developing dashboards to inform response efforts that overlay gridded population data with Covid-19 and socio-economic data to help governments identify vulnerable communities and map population risks.

Whereas in Southeast Asia, TReNDS’ expert member and former National Statistician of the Philippines, Lisa Bersales, stated that gridded population data has not yet been widely used for Covid-19 response. However, in 2013, the Philippine Statistics Authority (PSA) successfully used gridded population data to help respond to Typhoon Haiyan, and the PSA is considering implementing it again for Covid-19 efforts.

2. Understanding the tradeoffs of each dataset is important.

As demonstrated in our report, Leaving No One Off The Map: A Guide For Gridded Population Data For Sustainable Development, each gridded population dataset has its strengths and weaknesses, and these should be carefully considered before application. For instance, when choosing a dataset, it’s important to understand its input data, methodologies, and modeling approaches that are used to calculate the population estimate. TReNDS’ Hayden Dahmm noted that once you have a solid understanding of these nuances, “it’s recommended to then identify the different attributes for your particular use case and understand how these align with what’s available.” For example, if you are interested in where a population congregates during lunchtime, you should use a dataset that has daytime or 24-hour data. Additionally, if you opt for a high-resolution dataset, your results may be more prone to errors as you zoom in.

3. Gridded population data enhance census data.

A common misconception is that new data sources, such as gridded population data, can replace the census. However, most global gridded population datasets rely on census data to produce population estimates by using a top-down method to disaggregate or redistribute census-based population counts to grid cells. As Bersales stated, “Gridded population data are an enhancement to census data. It can paint a clearer picture of places that are difficult for enumerators to access, and in countries that don’t have regular funding for a census, it can help fill in the gaps.” Even in countries that do conduct regular censuses, gridded population data can provide more timely population projections in between census years. Especially during the current pandemic with countries facing challenges funding censuses and safely collecting data, gridded population data has become even more important.

Yet, as Bersales explained, most gridded population datasets use census data as part of their methodology so there can be challenges in areas that don’t have updated census data. Some countries also vary widely in the granularity of their data. For example, the U.S. has census data available at the census tract level (which average approximately 4,000 people), while in Russia, data is only available at the provincial level. These census gaps underscore the need to invest in robust statistical systems.

4. The aggregate form of gridded population data helps mitigate privacy concerns.

As more invasive technology solutions to tracking Covid-19 have emerged, questions surrounding data privacy and security are rightfully being raised. Fortunately, because gridded population data provide figures at an aggregate level - the most granular is approximately 100 meters x 100 meters – individual-level data are not available to users. As Dahmm clarified, “[gridded population data] are primarily looking at the community-level, rather than the individual-level, and this can help alleviate privacy concerns.” Like census data, it is collected at a more granular level, but is only released and analyzed in the aggregate form. Additionally, within these datasets, individual households have been anonymized in order to protect privacy.

5. Engagement with the scientific community is essential for making gridded population data more widely available.

The scientific community has an important role to play in improving the awareness and accessibility of these tools. As Victor Ohuruogu noted, in order for gridded population data to be more widely utilized, “there has to be effective engagement with the scientific community, researchers, and policymakers.” Lisa Bersales also highlighted the importance of use cases and the involvement of scientists in the policy-making process.

Groups like the International Science Council and the POPGRID Data Collaborative are helping to foster more dialogue between data users and researchers in the scientific community and to improve awareness and use of gridded population data for sustainable development.

You can watch a recording of the webinar here. For more information on gridded population data, read TReNDS’ new report.