Non-Intrusive Assessment of COVID-19 Lockdown Follow-Up and Impact Using Credit Card Information: Case Study in Chile

Area of Science:

• Data Science
• Urban Studies
• Epidemiology

Background:

• COVID-19 pandemic necessitated unprecedented public health interventions, including lockdowns.
• Assessing the real-world impact of these measures on population behavior is crucial.
• Traditional data collection methods for behavioral analysis can be intrusive and biased.

Purpose of the Study:

• To propose and validate a non-intrusive methodology for assessing lockdown impacts.
• To quantitatively analyze changes in population behavior using anonymized credit card data.
• To correlate economic activity shifts with lockdown enforcement in Santiago, Chile.

Main Methods:

• Utilized anonymized and geolocated credit card transaction data from Santiago, Chile.
• Employed unsupervised Latent Dirichlet Allocation (LDA) for semantic topic discovery.
• Analyzed temporal patterns in credit card usage to identify behavioral changes.

Main Results:

• Identified significant temporal patterns in credit card transaction activity.
• Observed a marked downturn (up to 70%) in economic activity in lockdown-affected communes.
• Confirmed behavioral changes using independent mobile phone connectivity data.
• Detected activity reduction preceding official lockdown enforcement.

Conclusions:

• The proposed methodology effectively assesses lockdown impacts using credit card data.
• Lockdowns significantly reduced economic activity, with effects observed before official mandates.
• Population behavior changes indicate proactive adoption of protective measures.