To get your sea legs with Caravel, we recommend starting with a sample data set.
Below we have a number of CSVs that contain reviews of popular apps. These data sources are refreshed nightly to contain recent reviews from the past six months.
Select a sample source from a category that is similar to yours below 👇
Download one of the sample CSVs above
Open the main side menu and navigate to Sources
Click New Source and choose CSV Upload as the source type
Give the new source a name (the name will show up in your sources filter in your feed), a description, and click Continue
Click Attach a .CSV file and navigate to where you downloaded the sample CSV. Select it.
In the import tab, select the properties of the CSV you want to evaluate
Under Select Text select the text columns you want to analyze. If you want to analyze more than one, click the + Add column button and select the additional columns.
If you'd like to be able to filter by the origin source and link back to the original review, be sure to map the proper columns under Origin.
Under Metadata be sure to select User name on the first drop down to associate the reviews with the user and Date under the second drop down to be able to filter and analyze the data by dates.
Click Finish to import the CSV.
The imported data will be available in Your Feed for analysis.
In the main side navigation, click Classifiers
Click New Classifier and select Pre-built as the training method
Pick a pre-built model from the menu. A good one to start with is Usability or Complaint.
Caravel will start analyzing all of your data with this new classifier. In less than a minute your classifier will be ready for use.
Once the classifier is done analyzing, navigate to Your Feed and select the name of your CSV source from the Sources dropdown.
Select your classifier from Classifiers dropdown to view the mentions in your text the classifier predicted