Delivery Format

Now that we've explained the dataset structure, we can go over format in which this data is delivered.

File Format

Due to the high number of variables, this is a large dataset. Only subsets of variables or geographies will be able to be opened in applications like Excel, and those opening this data in more advanced statistical or programming environments will often need to manage memory, especially when running on a local machine.

That said, this data is typically delivered in .csv format with each row representing a single individual or geography and each column representing a FollowGraph variable. Our standard geographic file is at the census block group level. If you are interested in other formats (H3 or census tract being some examples...), don't hesitate to ask your Spatial.ai representative.

The below table is an example table with 10 random selected variables (you can scroll the table right to see the whole thing) at a census block group level.

BLOCKGROUP	Boy Band Fans	Budget Fashion Enthusiasts	Business Readers	Dalai Lama	Daniel Dale	Olive Garden	Productivity Apps Enthusiasts	Rachael Ray Show	Soccer Enthusiasts	Cardi B
010010201001	119	141	60	74	60	146	58	126	85	92
010010201002	106	104	68	81	78	116	78	109	107	81
010010202001	123	165	76	83	76	151	71	125	92	149
010010202002	111	139	73	80	70	140	72	122	88	123
010010203001	105	125	72	78	75	131	75	120	93	104
...	...	...	...	...	...	...	...	...	...	...

All scores provided are index scores where 100 represents the national average. The first value in the first row is 119, indicating that the people in block group 010010201001 are 1.19 times as likely to follow Boy Bands on social media. This can also be stated as "19% more likely". This is further explained in the Dataset Structure section.

Your data file may include columns IDs instead of names (for the column headers). The "Key" file contains the full variable listing and you can use the IDs to match the data and key files. The key file contains a list of all the variables including names, descriptions, categorizations, and the social media accounts which each variable is based on.

Individual Level Data Delivery

For individual level data, the data looks different in the following ways:

1. The BLOCKGROUP column would simply be replaced by a Customer ID or similar ID field.
2. An additional column Match Level is included which details the level at which we have matched an individual. Possible values include (but are not limited to) Name, Address, Block Group, and Zip Code.

There's more info we'd like to emphasize on this, however, so we've devoted this entire documentation section to individual level data delivery.