Nomic Atlas Quickstart
Learn how to get started with Atlas to structure your unstructured data.
Here you’ll learn how to set up your Atlas account, start uploading your own datasets, and how to build your first map.
Make an Atlas account
Visit Nomic Atlas to sign up for a free account.
Set up your Nomic Atlas organization. Your Atlas Organization stores any datasets you upload and allows you to invite collaborators to your datasets.
Uploading your unstructured dataset
- Web Drag-and-Drop Upload
- API Upload
- Go to your Atlas dashboard and click the “New Dataset” button.
- Drag-and-drop your dataset in to create a new Atlas Dataset.
- Name your dataset and select the data field you are indexing.
- Optional: Configure settings on your dataset to disable/enable topic modeling, duplicate detection and initial access permissions. You should be good to go with the defaults.
- Time to index your dataset! Depending on the size and modality of your data, the indexing process may take anywhere from minutes for under 500k datapoints to an hour for multi-million point datasets.
Note: The field you select for indexing will determine how Atlas organizes your dataset. It's usually the unstructured data field in your dataset like a text column.
Those with knowledge of Python can upload data directly from their code. Walk through the steps below or jump directly into a 📓 Colab Notebook. Experienced developers may choose to go directly to our API reference or to our more advanced walkthroughs to get started.
Install Nomic package
- Install the
nomiclibrary in your Python environment. Virtual environments are recommended.
- Import the
pip install --upgrade nomic
from nomic import atlas
Log into Nomic
- Get your API key. You can obtain your API key at https://atlas.nomic.ai/cli-login. Make sure you already signed into Atlas in your browser at https://atlas.nomic.ai/
- Log-in to your Nomic account in your Python code (or your terminal) with your API token:
# in Python
# in the terminal
$ nomic login YOUR_API_TOKEN_HERE
Upload data and build map
- Read your data file and map it in Atlas. For additional parameters for
map_data, see the API reference.
- All done! Visit the map link to see the status of your map build. Read more to see what you can do with your map next.
from nomic import atlas
# Edit accordingly if working with other data type (TSV, JSON, JSONL)
with open('my_data.csv','r') as f:
reader = csv.DictReader(f)
my_data = [row for row in reader]
# Build a map using the map_data method
dataset = atlas.map_data(data=my_data,
description='A description of the data.',
dataset.maps # to view map build status