> ## Documentation Index
> Fetch the complete documentation index at: https://docs.getinteract.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Data Book

> Inspect the prepared dataset used to train your Meridian MMM model — column definitions, time range, spend series, and data quality.

The Data Book gives you a direct view into the BigQuery table that was used to train the MMM model. It's useful for verifying that the data is complete and correctly formatted before a model run, and for understanding what the model actually learned from.

## What it shows

When you open **Insights → MMM → Data Book**, Interact queries your configured BigQuery table and displays:

* **Time range** — first and last date in the dataset, and the number of observation periods
* **All columns** — one column per metric or channel variable, with values per date
* **Excluded columns** — columns listed in the "Excluded Columns" setting are hidden from the view

The data is shown as-is from BigQuery — no transformations are applied.

## Switching between input and final table

If you configured a **Final Table** in the MMM settings, a toggle appears at the top of the Data Book to switch between the raw input table and the adjusted/final version.

Use the final table view to verify post-processing steps (e.g. spend normalization or outlier corrections) before running the model.

## Multi-market configs

If you have more than one Data Book configured (e.g. NL and DE), a **label dropdown** appears at the top of the page. Select a market to switch which dataset is shown.

## Validating your data

Before running the Meridian pipeline, use the Data Book to check:

<Steps>
  <Step title="Date continuity">
    Confirm there are no unexpected gaps in the date sequence. Meridian expects a complete, contiguous time series.
  </Step>

  <Step title="Spend columns present">
    Every channel that will be modelled should appear as a column. Channels missing from the table cannot be included in the model.
  </Step>

  <Step title="KPI column present">
    The outcome variable (e.g. `revenue`, `conversions`) must be a column in the table.
  </Step>

  <Step title="No unexpected nulls">
    Scroll through the table and look for missing values. MMM models require complete data for all columns across the full time range.
  </Step>
</Steps>

## Settings reference

Data Book settings are configured in **Insights → Settings → MMM**.

| Setting                       | What it controls                                                           |
| ----------------------------- | -------------------------------------------------------------------------- |
| **Project / Dataset / Table** | Which BigQuery table to query for the Data Book                            |
| **Final Table**               | Optional alternate table (shown via the input/final toggle)                |
| **Date Column**               | Which column contains the observation date                                 |
| **Excluded Columns**          | Columns to hide from the Data Book view (e.g. internal IDs, country codes) |
| **Label**                     | Market or segment identifier shown in the dropdown                         |

<Card title="MMM setup" icon="gear" href="/insights/mmm/setup">
  Configure the BigQuery table and other settings.
</Card>