Importing Raw Data

Sometimes, you may have finished part of your experiment outside 3X, or your experiment result weren't produced by a computational process and you simply need 3X to visualize the data. You can follow the steps on this page to import a dataset directly into an existing 3X repository, or create a new repository from scratch based on the data. The dataset to import should be in CSV format.

Importing data into an existing repository

Suppose you have been studying the performance of sorting algorithms, and manually collected a few measurements in a file named quickSort-results.csv that looks like the following:

Algorithm,Input Size,Input Type,Sorting Time,#Compare,#Access
quickSort,15,random,0.390538,572959,1519407
quickSort,16,random,0.788945,1199432,2991392
quickSort,17,random,1.763706,2644539,6656851

Then, you can run the 3x index import-csv command with the name of the file you want to import, followed by the input or output parameter names that correspond to the columns in the CSV data (in the order of appearance, left to right).

3x index import-csv  quickSort-results.csv  algo inputSize inputType sortingTime numCompare numAccess

3X will add all the three rows of the data, and keep a copy of the CSV file in the repository.

Creating a new repository from a CSV file

Alternatively, if you have an existing experimental result dataset you simply want to explore using 3X, it is possible to create a new repository from a CSV file. As an example, let's say we want to browse the famous Fisher's Iris flower dataset that we have in a CSV file called iris.csv.

You can create a new 3X repository by importing the data in the file using the 3x import command. You need to pass the names of the parameters as arguments to the command, in the order that appears in the CSV data from left to right. Also, you need to indicate which are input and output parameters by preceding the names with --inputs and --outputs.

3x import iris-experiment iris.csv --outputs sepal_length sepal_width petal_length petal_width --inputs variety

3X will create a new repository named iris-experiment in the current working directory, define the experiment parameters, and import the data in the given CSV file. After it ends successfully, you can start the 3X GUI for the repository and browse the data.

cd iris-experiment
3x gui