• Description of Nighttime lights in Mexico at units of substantive interest repository
• Social science using night lights
• Acknowledgements
• Donate
• References

Last revision: 2023-12-30

News:

• 22nov2024 Municipio series consolidated in a single file here (by @jpfmansur Thank you!).

• 27mar2024 Possible update of the source data to 2021 here.

• 30may2021 Measures re-computed using INE's 2020 maps (instead of 2018 maps). Over-populated secciones are routinely split by INE, creating new units. Comments, critiques, and suggestions welcome by email.

• 5may2021 data/municipios/ and data/secciones/ now include luminosity measures for all 32 states. Validation detected a bug affecting smaller units (e.g. secciones in urban areas), producing large but short-lived surges/declines in luminosity measure. Bug has been fixed and data cleaned.

Description of Nighttime lights in Mexico at units of substantive interest repository

• Author: Eric Magar
• Location https://github.com/emagar/luminosity
• Email: emagar at itam dot mx
• Citation for the data: see 'About' on the repository landing page

The repository distributes nighttime luminosity data for Mexico, aggregated at the municipal (N approx. 2,500) and seccin electoral levels (N approx. 67,000). Luminosity rasters are from Li, Zhou, Zhao and Zhao's (2020) Harmonized Global Nighttime Light Dataset 1992-2018.

The source reports annual observations geo-coded at a spatial resolution of about one square kilometer (30 arc seconds). This repository summarizes pixels within each unit's territory, reporting the mean unit luminosity, the median, and the standard deviation. While the source averages (and harmonizes) pixel luminosity in time, this repo averages pixels in space. The chosen units have analytical potential, as they can be associated with census indicators, with electoral returns, and so forth. (Secciones electorales are Mexico's basic units for electoral management and cartography, analogous to U.S. census tracts.)

An example from one unit clarifies the summary statistics distributed. The unit in question is electoral seccin number 573 in Mexicali, the capital city of the state of Baja California. The following map shows the source's statewide data, shades of green indicating more nighttime luminosity, white less. Tijuana's brightness can be appreciated in the northwestern corner of the state, Ensenada's slightly southward, and Mexicali's eastward. The small red area is seccin 573, in the city's suburbs, towards the Colorado river delta (the river is the border with neighboring state of Sonora).

The map below zooms into the unit, portraying luminosity in seccin 573 and vincinity for the year 2018. (Shapefiles for secciones electorales and municipal boundaries are from INE's cartography department.) Mean artificial lights decreased near monotonically as one progressed towards the seccin's east-southeast. I relied on R's raster package (Hijmans 2019) in order to summarize the unit's luminosity.

The next map illustrates descriptive statistics generation (the code for this process is in script code/export-seccion-stats.r). R's raster::mask operation selects pixels inside the seccin's polygon, excluding the rest (which appear in white in the map). The mean luminosity, median, and standard deviation for the selected pixels are reported in annual state-by-state files in csv format inside folder data/secciones.

The time series can be appreciated in the final figure below. The seccin's rapid urbanization is notable. Most of the area was quite dark in the mid-1990s, with mean nighttime lights of roughly 8GW per pixel, tripling by 2018. The nighttime lightning aggregated in territorial units data offers a valuable indicator of human activity in Mexico's since the 1990s in small and analytically useful geographic units.

Social science using night lights

• Livny 2021 Can Religiosity Be Sensed with Satellite Data?
• Bruederle and Hodler 2018 Nighttime lights as a proxy for human development at the local level.
• Chen and Nordhaus 2011 Using luminosity data as a proxy for economic statistics.
• A critical piece by Pinkowski and Sala-i-Martin here.

• Growth reported relative to night lights by Martnez here.

• A Tweeter thread from Stanford AI Lab here.

• George and Ponattu 2018 Like father, like son? The economic impacts of political dynasties.

Acknowledgements

Eric Magar is grateful for financial support from the Asociacin Mexicana de Cultura A.C. He is responsible for mistakes and shortcomings in the data.

Donate

If you find this useful, help buy a cup of coffee for the author and research assistants. with Paypal.

References

• Hijmans, Robert J. 2019. raster: Geographic Data Analysis and Modeling ver 3.0-2, https://CRAN.R-project.org/package=raster.
• Li, Xuecao, Yuyu Zhou, Min Zhao, and Xia Zhao. 2020. A harmonized global nighttime light dataset 19922018. Scientific Data 7(1), https://doi.org/10.1038/s41597-020-0510-y.