We calculate GDP by transforming satellite imagery using deep learning

Transforming night lights from satellite imagery to GDP was pioneered by a LSE professor from our Department in 2012 in the American Economic Review, the world’s top economic journal. Since then, thousands of papers of peer-reviewed work has been published on transforming luminosity to GDP, all of which find a very strong relationship between these two variables. We further improve on the literature by combining traditional academic approaches with our proprietary machine learning algorithms.

uklightsukgdps
1
2
3

First appearance of luminosity and GDP in academia (2012)

Night light luminosity data taken from satellite imagery was first used as a proxy for GDP in the American Economic Review (world’s top Economic journal), Measuring Economic Growth from Outer Space (2012), authored by Professor Vernon Henderson et al., from the LSE Department of Geography and Environment, the #1 ranked Department in Economic Geography globally.

Peer-reviewed scrutiny (2012-2020)

1000s of papers published on luminosity and GDP.

Consistently find 70-90% correlation between luminosity and GDP from national statistics

Luminosity is one of the most rigorously scrutinized proxies for GDP

505’s methods (July 2019 – now)

Improvements over traditional literature include higher resolution satellite imagery, a different way of processing raw imagery, and deep learning algorithms to transform luminosity to GDP.

At a national-level, our estimates differ from official statistics by ~2%. Our team is comprised of PhDs and postdocs, who have published peer-reviewed papers related to transforming luminosity to GDP.

First appearance of luminosity and GDP in academia (2012)

Night light luminosity data taken from satellite imagery was first used as a proxy for GDP in the American Economic Review (world’s top Economic journal), Measuring Economic Growth from Outer Space (2012), authored by Professor Vernon Henderson et al., from the LSE Department of Geography and Environment, the #1 ranked Department in Economic Geography globally.

1

Peer-reviewed scrutiny (2012-2020)

1000s of papers published on luminosity and GDP.

Consistently find 70-90% correlation between luminosity and GDP from national statistics

Luminosity is one of the most rigorously scrutinized proxies for GDP

2

505’s methods (July 2019 – now)

Improvements over traditional literature include higher resolution satellite imagery, a different way of processing raw imagery, and deep learning algorithms to transform luminosity to GDP.

At a national-level, our estimates differ from official statistics by ~2%. Our team is comprised of PhDs and postdocs, who have published peer-reviewed papers related to transforming luminosity to GDP.

3

A luminosity-based approach provides a robust foundation for proxying GDP

Other indicators, like mobile location data or search trends data, can be layered on top when necessary. 

Luminosity data has global coverage, high granularity, and is comparable across time.

More details

Luminosity data has a much longer time series than other indicators, like mobile phone location data which only really exists in recent years.

Luminosity data is calculated consistently across geographies, compared to other indicators that might have more skewed samples (e.g. mobile location data/ credit card transaction data). 

  1. Electrification programs in developing countries can introduce bias

    In developing countries, governments may introduce electrification programs. Our model may interpret this as a shock that increases GDP.

    This is only limited to more remote areas in less developed countries (e.g. cities generally have gradual electrifications).

  2. Cloud cover can block luminosity

    We clean images by removing cloud cover using code. However, some countries (e.g. Northern countries in Scandinavia) experience months with cloud cover, which makes this process more difficult.

Some customers who have found our data useful include MBB consulting firm(s) and multi-billion dollar ride-sharing companies.

Our team is comprised of PhDs and postdocs from the #1 ranked Department in Economic Geography globally, who have published peer-reviewed papers related to transforming luminosity to GDP, and won awards like the Forbes 30 under 30 in data-related initiatives.

Test if our data is suitable for your use case

Compare regions’ GDP data in the year 2015 using our free visualization tool

Download all our subnational  data  for 1000+ regions in 2015, broken down monthly, for free