AI App Catalogue

One-shot Semantic Segmentation

This Dash app uses machine learning in order to compute the segmentation of an image, given user-provided annotations. In addition to Plotly and Dash, the app uses off-the-shelf algorithms and estimators from PyData packages, namely scikit-image and scikit-learn. 

Access this Dash app and get the Python code. 

t-SNE Explorer

T-SNE is a visualization algorithm that projects your high-dimensional data into a 2D or 3D space so that you can explore the spatial distribution of your data.This t-SNE Explorer lets you interactively explore iconic image datasets such as MNIST, and state-of-the-art word embeddings such as GloVe, with all the computation done ahead of time. 

This app's source code is available in Python and R.

Handwriting recognition

This app uses Optical Character Recognition (OCR) to convert handwriting into text.

Source code

Mnist Explorer

This app applies T-SNE on the images from the fashion mnist dataset, which reduces each image to a two dimensional embedding to visualize the similarity between images. Clusters represent similar images, and the greater the distance between two points, the less similar the images are. The app also allows to predict the class of each image using a simple convolutional neural network (CNN). The T-SNE embeddings were generated using RAPIDS cuML, and the CNN was trained using Keras.

App is available in Python.

Basketball Player Pose Estimation

Learn more about this app in our post: 7 Dash Apps Bringing AI & ML to Sports Analytics.

Reach out to us for the source code.

Generative Adversarial Network (GAN)

GAN is a type of machine learning model that uses a generator and discriminator neural network to generate outputs.

We used GAN and PyTorch to produce a Dash app that generates realistic face images based on a selection of user-chosen physical features. Read more about how to build apps for editing Face GANs with Dash and Pytorch Hub.

Try the app!

3D Brain Annotation 

This app allows you to:

• Interactively move through sliced representations of the brain
• See an overlay of the automatically computed over-segmentation of the brain into superpixels (regions of the brain)
• Highlight one or several superpixels
• See the highlighted superpixels rendered in a 3D glass brain representation
• Download a file with that contains the highlighted superpixels 

Access this Dash app and get the Python code. 

Customer Complaints Analysis 

This Dash app performs natural language processing on a database of consumer complaints towards US financial institutions. Compare consumer complaint sentiments between Wells Fargo, Citibank, Equifax, JP Morgan and dozens of other institutions. 

Access this Dash app and get the Python code. 

Product Reviews Insights

Snowflake is a cloud-only, distributed commercial data warehouse with drivers for Python and R.

BI tools like PowerBI or Tableau are typical front ends for Snowflake — use Dash when you need an AI front end for NLP, computer vision, predictive analytics, or deep learning using data stored in Snowflake.

This Dash app performs NLP on product reviews stored in Snowflake - all in less than 1,000 lines of Python.

Word Embedding Arithmetics

Dash is the fastest way to deploy Python-based apps for natural language processing (NLP). 

Use NLP + Dash to extract business value from large amounts of unstructured text:

• Digitized company documents
• Research articles
• Customer reviews
• Transcribed audio recordings
• Social media streams

This Dash app demos Google's word2vec in ~300 lines of Python code

Named Entity Recognition

An NLP app that uses Spacy to perform named entity recognition, trained on Wikipedia and OntoNotes datasets.

This application was built using Dash Enterprise Design Kit, which lets you easily arrange, style and customize Dash apps so that they can be deployed in no time. 

Interest Rate Modeling 

This Dash lets you train and evaluate a L2 regression (Ridge) model for predicting the interest rates of loans based on various factors. It lets you see how well a model performs (actual vs predicted plot), as well as compare the importance of these features. To read the model data, this Dash app connects to a Snowflake data warehouse. 

Access this Dash app and get the Python code. 

Sales Forecasting   

This Dash app uses Facebook's Prophet time series forecasting model to perform predictive analysis of regional sales. The data was collected for locations across various states and for various market segments, and the model was used to forecast future sales, which is useful for making informed decision and infer the supply needed for future months. 

Access this Dash app and get the Python code.