Lang.ai allows you to create a project and then upload new data to take advantage of your project's tags and Dashboard. This means that all the new data will be already classified and searchable.
1) Create a project starting from a dataset
Click on Create Project > Start from a Dataset to get started.
The supported file formats are:
- CSV files
- Plain text files, separated by line breaks
In order to take full advantage of the Dashboard section, you will need to upload a CSV file with a valid date column. This way you will be able to filter your tags and documents by date, making it easy to get insights about trends.
In case you have an issue creating a project, make sure that your documents are encoded in UTF-8.
Keep in mind that all the CSV columns will be available as metadata filters as well. For example, if you include a column like "Country", you will be able to filter by it in inside your project. You can configure the fields available in the Setup > Metadata section of your project.
2) Upload new datasets to an existing project
Once a project is created, you can upload new datasets to take advantage of the project's classifier. This means that the new data will be tagged (if your project has at least one tag) and classified with intents. In the Projects lists, click on Select an Action > Upload new dataset.
For example, you can start by creating a project to analyze customer support reviews with a dataset from the previous quarter. Once the project is ready, you can upload the reviews from the current quarter and get them tagged and classified to get insights on what changed, what topics increased, among others.
3) Filter your project's documents by date
Once your new documents are uploaded in your project, you can filter by date in both the Review and Dashboards sections in order to get the insights you are looking for.
Filtering by date will update the Tags sidebar and the Tags visualization.
Filtering by date will update all the blocks from this section: Evolution, Distribution, Discovery, and the charts that show an aggregation by each metadata field activated.
You can check our FAQ about datasets here.