Analysis Guide
In the Analysis section, you can analyze your own text (up to 4000 characters), choose a random quote, or upload a CSV file containing your dataset.
If you choose the Text path:
- First, select a predefined topical model or your own text model (if activated). Clicking on a predefined model will display more information about it underneath.
- Next, insert your own text in the text field area or choose from ready-made example sentences. Once the analysis is finished, you will see a results page with five tabs.
TIP: In the text preview window seen on the first four tabs, you can hover over a word to see a small tooltip with the word's unified base and type!
-
CLASSIFICATION- In the table, you'll find the model used for processing your sentence, the category, and the result recognized by the applied model. The "Match A and B" section provides an explanation of the result.
-
CLAUSES- Here, your sentence(s) are broken down into smaller parts called clauses.
-
PHRASES- Here, your sentence(s) are broken down into their constituent parts and classified into various grammatical phrases to help you understand their structure and relationships within the sentences.
-
GRAMMAR- Here, the grammatical structures of your sentence(s) are identified. Each relationship describes how one word or phrase is connected to another in terms of grammatical structure and semantic meaning. This tab divides phrases into verb and noun phrases and assigns roles to each word or phrase. Underneath the field with your text, you can find the detected language.
-
JSON- Here, you can find a structured, machine-readable output of the entire text analysis in JSON format.
If you choose the CSV file path:
- Upload your CSV file containing the dataset. Ensure that your file complies with the supported standard. You can upload it by dragging the file to the drop area or click on the
BROWSE...button and select a file. Remember that in this version, you'll see results for the first 100 records.
As your file is imported successfully, on the right bottom you will see a notification:
- As your file is uploaded, select a column with a unique ID (step 1) and a column with text for analysis (step 2). If you need to review the original file, use the
Preview CSVtab above. In Step 3, you can choose any columns from the original file that you wish to include in the results file. Then, click on theLoadbutton.
- Next, select a predefined topical model or your own text model (if activated) and click the
Analyzebutton to start.
As the analysis is processing, you will see a progress screen displaying the number of successfully analyzed texts, failed analyses, and timeouts. If you click on the Cancel button during the process, you will see the results for the texts that have been analyzed until that time.
Once the analysis process is finished, you will see three buttons:
-
Raw Data- It shows how your text is decomposed into its grammatical components and the relationships between different parts of it. It provides a clear and structured view of the raw data, facilitating deeper analysis and insight into the linguistic features of your text.
-
Results- It shows your data processed and analyzed. You can find text features as an output of the activated model processing. At the top of the window, there are options to export the results in either CSV or JSON format, allowing for further analysis and data usage outside of the Smabbler Portal. It provides the high-level outcomes of your text analysis, showing how each record in your dataset has been processed and classified according to the selected model.
-
Visualizations- Here you can view your classified data in pie and bar chart formats.
To ensure compatibility and successful analysis, your CSV files must adhere to the following standard:
-
Delimiter: Semicolon (
;) - Encoding: UTF-8
- Maximum file size: 10 MB