# DATASET EXAMPLES

Question. What types of datasets can be analyzed with 'Mparanza?

Answer. Flat sales datasets. 'Mparanza can also analyze flat cost datasets with a similar structure.

You can feed 'Mparanza sell-out POS datasets such as those provided by the likes of Nielsen and IRI to understand what is going on in a market or specifically to a company or a brand.

You can feed it your sales actual and actual vs budget data. Add indirect costs and the app will return an "EBITDA bridge" to illuminate how your net margin has changed.

The app offers a number of example datasets you can play with. Select one with the "Choose an example📁dataset" widget.

The app requires the dataset to have a flat "tidy" structure, one row per observation.

The date and metric columns must have a defined naming convention. Beyond these requirements, the app is relatively agnostic and accepts a wide variety of different dataset setups.

Here are a few examples.

##### 1. Period column, units, amount, discounts, cogs

A dataset with this setup will return the sales and the gross margin variance of the two periods, with price - volume - mix split. It will separately detail the impact on variance of changes in unit discount from the impact on variance of changes in unit cost.

##### 2. Period column, units, amount, discounts, cogs, indirect costs.

A dataset with this setup will return the same as 1 above. Since there is an indirect costs column, it will return net margin variance, not gross margin variance, building an EBITDA bridge.

##### 3. Period column, units, amount, cogs, indirect costs.

A dataset with this setup will return the same as 1 above, but will not detail the impact of changes in unit discount since it have not discount column. Since there is an indirect costs column, it will return net margin variance, not gross margin variance, building an EBITDA bridge.

##### 4. Period column, units, amount, discounts, cogs, visits, conversions

A dataset with this setup will return same as 1 above, plus the impact of visits and conversion drivers on sales variance.

##### 5. Period column, units, amount, discounts, weighted distribution

A dataset with this setup will return same as 1 above, but not gross margin variance. It will also return the impact on sales variance of the change of distribution.

##### 6. Period column, units, amount, volume, weighted distribution

Same as 5 but also includes volumes (in this example weight) column.

##### 7. Period column, units (defects), amount (downtime)

To analyze supplier quality analysis. The units column maps the number of defects. The amount column maps the downtime generated by the defects. Price maps the gravity of the defect.

##### 8. Period column, amount

A dataset with this setup will return the aggregated sales or cost variance column without the price-volume split since there is no unit column.

##### 9. Period column, units (investment amount), amount (revenues or costs)

To calculate the ROI variance of a set of initiatives between two periods. The amount column maps the revenues (with positive sign) and the costs (with negative sign) of a given initiative. The unit column maps the investment cost of the initiative (in positive). The "metric" column (that does not need to be named "metric" ) is a dimension column that simply describes the nature of the revenues and costs.

##### 10. Scenario (Plan/Actual) column, units, amount, cogs

A dataset with this setup will return same as 1 above. Sales and margin variance are not measured between two periods but between actual and plan scenarios, with price - volume - mix split. It will not detail the impact of changes in unit discount since there is no discount column.

##### 11. Scenario (Plan/Actual) column, date column, unit, amount, cogs

A dataset with this setup will return same as 8 above. The date column makes it possible to compare actual and plan across different intervals.

##### 12. Scenario (Plan/Actual) column, date column, amount, cogs

A dataset with this setup will return same as 8 above, excluding price-volume split, since there is no unit column. The date column makes it possible to compare actual and plan across different intervals.