A CSV series of vehicle counts from satellite imagery of the chain's parking lots, seasonally adjusted, updated daily.
01What you get
| File | Format | What's in it | Cadence |
|---|---|---|---|
| Chain index | CSV | One row per day for the chain: the vehicle count summed across the lots in scope, the indexed value and the seasonally adjusted value. Updated daily. It loads straight into the model you already keep for the name. | Daily |
| Store-level series | CSV, optional | The same series broken out by lot, keyed by a store identifier from the list you supplied. Use it to check which locations move the chain figure, or to drop a store that is closed or under construction. | Daily |
| Coverage flags | CSV column | A flag on each daily value where cloud cover or a capture gap meant a lot was missing or counted with lower confidence, so your model can drop or down-weight the day instead of treating it as a real zero. | Per daily value |
| Store list | CSV | The chain estate as scoped: each store, its lot, and whether it is in the chain figure. It is the file you check the index against. | As scoped |
02Pipeline
01Input
Tell us the chain and, if you have it, a store list with addresses. We confirm which lots have visible surface parking.
02Imagery
Very high resolution satellite or aerial imagery is captured over each lot in scope, on the days it is available.
03Counting
Vehicles are counted in each lot. The counts are summed to the chain, indexed against a base period and seasonally adjusted.
04Output
You get the series as a running CSV, chain-level with a store breakout if you asked for one, ready to join to your own model.
03Operating envelope
These are where a lot count stops being a sales read. They belong in your model as much as on this page.
| Condition | Effect on the series |
|---|---|
| Cars, not sales | The series counts vehicles in a lot. It says nothing about basket size, price, mix or what shoppers bought, and it is one input to a nowcast, not a forecast of the print. |
| In-store visits only | Online orders, delivery and pickup that never touches the lot do not show. The signal is weaker for online-heavy and delivery-heavy retailers and fits chains where physical visits still track sales. |
| Cloud and capture gaps | A cloudy day, or a pass that missed a lot, can leave a missing or lower-confidence count for that date. Those days are flagged in the file, not filled in. |
| Daily is a target | Daily is what the series aims for. Imagery is captured when it is available, so some lots will have days without a usable capture. |
| What ≤0.5 m resolves | At ≤0.5 m a car is only a few pixels across. That is enough to count vehicles in a lot, but not to tell models apart, and cars parked close together or under trees, canopies or a shadow can be undercounted. |
| Surface lots only | Multi-level garages and covered parking are out of view. A chain with mostly structured parking is a poor fit. |
| Other reasons a lot is full | A lot can fill for an event, a closure next door or roadworks. Store-level series help you spot it, but the series cannot tell you why a lot was full. |
04The job today
Before the print you have management's last call, card-panel data, an app-based footfall panel and a few channel checks. The panel sees only the phones that carry the app, and a store visit is one lot on one afternoon.
In the final weeks of the quarter, when the model needs a read on whether stores are busier than last year, there is nothing that covers the chain estate. This puts a count on every lot you name, every day the imagery allows.
05Fit
Exclusions
06FAQ
No. It is an indexed traffic series, one input to your nowcast alongside card-panel and app data. We make no claim that it tracks same-store sales for any given quarter, and we give no backtest figure.
Any chain with visible surface parking, listed or not. It fits chains where in-store visits still drive sales and fits poorly where online or delivery is a large part of the business.
That depends on the chain and the lots. Ask about your names and we tell you what is available before you agree anything.
As missing or lower-confidence values, flagged on the date. We do not interpolate across a gap.
Line the series up against the quarters the chain has already reported and against your other panels. Tell us which periods you want it tested on when the chain is scoped.
Yes, as an optional breakout keyed to the store list you supply.
It is a CSV, so it loads into Excel, Python or R, or the data platform you already run. Tell us the schema you expect and we tell you what fits.
Use and redistribution terms for the series are set in the contract before delivery. Talk to us about whether it can be used internally only or shared in published research.
Each chain is scoped and quoted on its own, from the number of stores and the history you want. You get a written scope for your procurement or compliance review.
Or ask about a specific name .
Get started