# Vignette 15: Mexico's positive deviance
In Mexico City, where over two-thirds of women have experienced violence and around 90% of incidents go unreported, the UNDP Accelerator Lab sought to identify and understand public spaces that, despite sharing similar characteristics with more dangerous areas, were unexpectedly safer for women.[\[1\]](#endnote-1) Over more than two years, from March 2020 to June 2022, the team undertook this data-powered positive deviance study.
The team used the Data Powered Positive Deviance[\[2\]](#endnote-2) method to analyze crime data across 2,414 basic geostatistical areas (AGEBs) in Mexico City. "Instead of focusing only on symptoms, we mapped the entire system to understand deeper patterns," explains Gabriela Ríos, Head of Exploration at UNDP Mexico's Accelerator Lab.
Working with Codeando México, they combined diverse data sources, from investigation files and urban infrastructure data to socioeconomic indicators. They categorized crimes by severity and used statistical modeling to identify areas with lower crime rates than predicted. Through clustering analysis, they examined how factors like population density, marginalization indices, and daily commuter patterns affected women's safety.
A breakthrough came when the analysis revealed that some areas with characteristics typically associated with higher crime rates, like informal commerce, financial services, and bars, actually showed better safety outcomes. This suggested there were hidden factors making these spaces safer that couldn't be explained by traditional assumptions.
To validate these statistical findings and understand factors that couldn't be captured in quantitative data, the team designed a qualitative research phase. They used a range of ethnographic techniques, including participant observation, feminist participatory cartography (Figure 30), and exploratory walks. Working with Cohesión Comunitaria e Innovación Social A.C., they investigated how the occupation of public spaces, interactions between people, and perceptions of safety contributed to women's actual safety in 16 carefully selected AGEBs.
The project established new ways of understanding safe public spaces by leveraging positive deviance through a combination of quantitative data analysis with deep community insights.. The findings are being used by Mexico City's Ministry of Women, Ministry of Citizen Security, and other government agencies to inform policy decisions and urban interventions that can make public spaces safer for women.
{% hint style="info" %}
#### **Key takeaways:**
* **Combine unusual and traditional data sources:** Mix investigation files, urban infrastructure data, and socioeconomic indicators to reveal patterns invisible through single data streams
* **Dig deeper into surprising findings:** When areas with "risk factors" like bars and informal commerce show better outcomes, investigate what's really happening behind these counterintuitive patterns
* **Let the data guide your fieldwork:** Use statistical analysis to identify which places are worth visiting and what specific questions to ask when you get there
* **Choose research methods that fit your context:** Use approaches like community mapping that can capture people's real experiences in ways your data cannot
* **Build partnerships that span the research-to-policy gap:** Engage government agencies from the start to ensure findings translate into actionable urban interventions and policy decisions.
{% endhint %}
***
## Notes
1. Also see Cervantes, Ríos, and Soto (2021); Rios Landa (2022). [↑](#endnote-ref-1)
2. UNDP, GIZ Data Lab, & University of Manchester (2021). [↑](#endnote-ref-2)
---
# Agent Instructions: Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.
Perform an HTTP GET request on the current page URL with the `ask` query parameter:
```
GET https://bas-leurss-personal-organization.gitbook.io/undp-accelerator-labs/doing-r-and-d/5.-r-and-d-practices/analyzing-data-and-generating-insights/vignette-15-mexicos-positive-deviance.md?ask=
```
The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.
Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.