CHECK-IN: From raw data to time series

Creating time series from CSV data – Step by step

Here’s a compact guide that walks you through all the steps to create time series from your CSV files in the futureNOW frontend. The video above also takes you through each step, one by one.

The key steps at a glance

• Upload data and start CHECK-IN
• Step 1: Delete unnecessary data
• Step 2: Define data structure
• Step 3: Check and clean data
• Step 4: Create time series
• Using the results
• More Resources

Upload data and start CHECK-IN (Minute 0:00 - 0:40)

First things first – log in to the futureNOW frontend.

1. In the left-hand menu, go to Data.
2. Upload your dataset:
   • Click on the plus icon.
   • Drag and drop your CSV file or select it manually.
3. Start the CHECK-IN by clicking the corresponding button next to your CSV file.

Step 1: Delete unnecessary data (Minute 0:40 - 0:59)

Remove any rows or columns that aren’t needed for the forecasts.

1. Click on the rows or columns you want to delete.
2. Confirm deletion by clicking the trash bin icon.

Step 2: Define data structure (Minute 0:59 - 1:42)

Now define the structure of your CSV file.

1. Select:
   • Decimal separator and thousands separator.
   • The strings used for missing values.
2. Specify:
   • The column containing date information and its date format.
   • The columns that contain the values.
   • Any grouping variables.

Step 3: Check and clean data (Minute 1:42 - 2:26)

The automatic plausibility check highlights all faulty or unclear cases. Plus, you’ll get descriptive stats and visualizations to give you a quick overview of your raw data. That way, you can check and clean your data easily.

1. Left side: Identify issues (e.g. invalid formats, missing values) in your data.
   • Click the exclamation mark next to each column to see specific problems.
   • Remove faulty rows or enter correction values.
2. Right side: Explore various visualizations to get a feel for your data.

Step 4: Create time series (Minute 2:26 - 3:10)

Define how your time series should be created.

1. Choose the desired temporal granularity.
2. Optionally set a start and end date.
3. Select the grouping columns as the data level for your time series.
4. Set up include/exclude filters for individual groups.
5. Specify:
   • The value variable.
   • The type of aggregation.
   • Whether to fill in missing data points with 0.
6. Click on Prepare time series and give it a moment.

Using the results (Minute 3:10 - 3:30)

1. The processed time series now appear in the table.
2. You can select individual time series from the table to view them. A chart on the side will then show the historical trend (e.g. “Material”).
3. Export the results as a CSV or use them directly to create forecasts. For that, we recommend checking out our “Quickstart” guide.

More Resources

Still got questions? Here are some helpful links that will support you in making CHECK-IN part of your workflows.

• Once you’re logged in, the frontend futureNOW also has a detailed manual for CHECK-IN and all other features of futureNOW.
• You can also start the CHECK-IN directly using the Python client. Check out all the info in our GitHub Notebook: EXPERT CHECK-IN options
• Our blog post on Demand Planning shows how we use CHECK-IN to create tailored forecasts for demand planning.

Do you still have questions, ideas or remarks?
Then feel free to contact us by email at any time: support@future-forecasting.de or simply create an issue on GitHub for new features.
With your input we can make future even better!

We’re looking forward to you!