Creating a custom recipe inside our Custom Recipe Creator is very simple, you will just need to answer a few questions regarding the kind of data you are working with.

1. ANALYSIS IDENTIFICATION

The first block of questions tries to identify the data you are working with. You will now need to choose between these data kinds:

  • Customers
  • Surveyed people
  • Employees
  • Tweets
  • Images
  • Text
  • Others

Depending on the answer to the previous question you may need to answer another question narrowing down the purpose of the analysis.

2. DATA MAPPING

The second block of questions asks you to identify the columns of the dataset you are working with. Depending on the usecase you may be asked about:

  • Your target column: Column you want to explain, it will be excluded for similarity.
  • Columns to exclude in similarity: Usually IDs, Emails, categoricals with a very high cardinalyty... These columns that are marked as not relevant wont be used to define the proximity of nodes in the network.
  • Columns to use for similarity: The selected columns will be used to define the proximity of nodes in the network.

Recipes on image analysis will require you to provide:

Recipes on text analysis will require you to provide:

  • Column containg the url to the text or the text itself.

Recipes on tweet analysis may require you to provide:

  • Column containg the tweet id.
  • Column containing the twitter handler of the writer.
  • Column containing the url or content of the tweet.

3. DATA ENRICHMENT

The next block of questions suggests you different enrichment options based on the kind of analysis you are doing. The main enrichment options available are:

  • Census data: you will need to provide an address or geographical coordinates.
  • Contact info using fullcontact: you will need to provide a column containing emails and a full contact API key
  • Gender: you will need to provide a column containing first names.

If you do not wish to enrich your information with any of the above you can simply select the "do not enrich" option.

4. DATA VISUALIZATION

The last block of questions configures the visualization of your data. You will be asked about the column you wish to use to identify your data. It will appear as labels above the nodes in your network, this is how it looks like:

5. NAME PROJECT AND EXECUTE

Name project and press “EXECUTE” and project will start creating. Please note projects usually take a couple of minutes to be created.

Did this answer your question?