Your “like” does matter: Using social media data to strengthen UNDP Peru’s impact

By UNDP Peru Accelerator Labs

24 de Enero de 2023


Most of us has an active profile on some social network: Facebook, Instagram, Twitter, TikTok, or LinkedIn, just to name a few. We communicate, learn, find out about current events and entertain ourselves. But how could social media help close development gaps? Do social media have space within sustainable development at all? In this blog post we present an experiment where the UNDP Peru Accelerator Lab worked used social listening to widen our view on citizens’ needs and how to respond to them.

Our experiment with social listening began in early 2021, a few months after the Lab started operating and also a few months after the most widespread protest in Peruvian history in November 2020. Young people took the streets, TikTok influencers who usually shared lifestyle tips showed how they stole their mothers’ pots to bang them outside their windows, and graphic designers shared downloadable files that could be printed and used as signs in protests. UNDP was well aware of the importance of social media as tools to articulate citizen proposals and spread knowledge, but it became clear that we had to strengthen our internal capabilities to process the immense amount of information from social media and understand it beyond our own echo chambers.

What is an echo chamber? It’s a digital literacy concept that explains how social media algorithms tend to show us content that reflects our own views, preferences and interests. In other words, we “hear” an echo of the content we produce or seek in the first place, and unless we intentionally look for it, it’s unlikely that we’ll read about or hear ideas that aren’t similar to ours. It’s not so much about being right or wrong, but about being aware of these biases and how they can influence project design and implementation to make sure no one is left behind. 




Understanding the internal user’s needs

UNDP has several partnerships with platforms that send alerts when a post it shared more often that usual such as RapidMiner or allow the use of public data for development projects such as Facebook Data for Good, but what kind of platform or service works best to address UNDP Peru’s needs? This was the question that guided some of the first insights in this experiment, and the answers were essential for the procurement process. Some of our requirements were:

  • We needed the option to access data in real time in case of crisis or a specific inquiry, but we didn’t have enough resources (time or a dedicated person) to analyze large quantities of data produced weekly and monthly;
  • Data from different social media on a certain subject had to understood holistically, including demographic data from users;
  • We needed a platform that learned from our interactions through machine learning;
  • We required historic data (from one year back) that allowed us to track how conversations evolved through time;
  • Training sessions for colleagues from different teams within UNDP were non-negotiable, as many as necessary. 

With these clear needs we kickstarted our procurement process, which led us to our second learning: There is a gap between the Lab’s experiment format that ranges from 3 to 6 months and the minimum time requirement that SaaS providers ask to submit a tender offer. The Lab wanted to hire the services of a social listening provider that could fulfill all internal requirements listed above and give regular in-depth reports for specific topics, but only for four months since it was still an experiment. This bidding process was declared deserted because service providers need contracts that last for at least one year, so we decided to extend the experiment’s timeframe to a year to adapt to the existing market conditions and still ensure that the service met all technical requirements. 


The day after procurement: The rollout

Once the procurement process was over, the internal work began: How could we show the benefits and possibilities (and also limitations)  of the service we were about to rollout and have access to for a year? The solution we found was to set up dashboards to monitor keywords or short phrases provided by each programme that were relevant to UNDP projects (for example “congress”, “elections”, “informality” or “digitalization”), and through these dashboards understand our colleagues’ needs according to the political and social context in which their projects exist. Each programme officer selected focal points in their teams that had editing rights for their programmes’ dashboards, in addition to receiving a weekly email with highlights. Besides collecting and organizing these topics, colleagues from all teams received over 10 hours of training in total, including sessions for specific teams such as the Governance Programme’s RedPublica team or colleagues from the Resident Coordinator’s Office. 

This tool allowed us to understand the political and social context from the perspective of social media and break out of our own “echo chambers”, while also helping us identify who is influencing or leading public debate on topics of our interest. Keep in mind that all data analyzed is public data, so no user’s privacy was impacted at any point. We were also able to understand the impact that some events or campaigns organized by UNDP Peru, such as RedPublica or the National Dialogues on Food Systems, or national occurences, like the January 2022 oil spill on the Peruvian coastline, had on social media. 



Furthermore, through social listening we have been able to validate insights found in project design or piloting stages. For example, in 2022 we piloted CREANDO, a project that seeks to strengthen entrepreneurial skills among Venezuelan migrants and refugees and the host population, and that will scale to in-person trainings and events. Beyond positive feedback we received from participants, we have used social listening data to confirm the project’s content structure and the migrant population’s interest in this type of learning experience.



A final note

A final learning, and likely the most important one: Having an additional data source enriches our perspective and lets us open the lens to escape our own biases, that in turn are strengthened by social media’s algorithm; however, having enough time and processes to internalize and apply new insights is just as important as having the data itself. Is there someone in your team with this responsibility and enough time to do it? Do you have a process to incorporate insights derived from social media data? Could we use this data on a regular basis or do we react only when there’s a crisis?


We want to meet you and learn from you! Email us at and tell us about new data sources that you’re using and how you’ve managed (or not) to integrate them into the established processes in your organization.