Frontier — An Analytics Workbench by FluidAI
Artificial Intelligence (AI) has grown leaps and bounds over the past few years and there is little doubt that it is only going to get bigger and better from here. It relies on various techniques and algorithms to generate insights, patterns and relationships in the data. Data can be very comprehensive to start with and requires proper handling in order to extract accurate and valuable information. We understand that in can be a little intimidating to start with. Thus we at FluidAI, believe that nobody should be hesitant of using it and instead everyone should be aware of the power that AI brings with itself. After all, we all want our businesses to grow beyond our imagination and FluidAI believes everyone should have the opportunity to make full use of their data and resources.
Data Science is the branch of AI which deals with algorithms, methods and processes to gather knowledge from unstructured data. To obtain such an information a certain lifecycle has to be followed such as:
- Data Fetching
- Data Cleaning
- Data Visualization
- Feature Extraction
- Feature Engineering
- Model Training
For an experienced Data Scientist all these steps might seem normal but for others this might be a herculean task to perform them. We at FluidAI understand this very well and we bring you a platform which does all the above tasks in an organized manner and provides an interactive way of working with Data. In addition, we follow a pro-active End-to-End Automation Model where the Predictions (ex. delinquencies, defaulters) are notified in advance. The Collections Platform automates the methodology and frequency of communication thereby reducing the last minute costs which could prove to be a setback during collections. It covers a wide range of channels for communication such as text messages, e-mails, Slack, WhatsApp messages or even AI voice calls.
Frontier is a platform designed for people from different fields who want to use AI to drive their businesses to greater heights. Thus, it serves as the quintessential AI platform covering all the necessary features required.
- Environment Friendly: Frontier can be deployed in Cloud or as an on-premise server offering flexibility to organizations all across.
- Projects and Versioning: This feature allows the user to create multiple projects and have multiple versions of Model in it.
- Pipelines: Pipelines are predefined data flows to ensure a particular outcome to User. Frontier allows creation of multiple pipelines for a specific project and there are a few important ones which come with the product itself such as:
- Database Pipelines: Users can connect to various databases such MySQL, MsSQL, Oracle, Snowflake and MongoDB in just a few steps. Using SQL statements data can be fetched from the databases directly and stored as csv files.
- Data Vectorization: Data as we know requires Feature Engineering to be fed to AI models, frontier also takes care of that. Various techniques such as Standardization, Normalization and One-Hot Encoding are available at your fingertips to make things easier.
- Model Training: Model Training is now easier than ever as one can select the training features and the target variable after vectorization with choice of appropriate algorithm and BINGO!.
Oh! and we have AutoML pipeline as well if you don’t want to wrap your head around choosing which algorithm.
- Reports and Visualizations: Not all of us want to generate predictions without knowing how the Data looks like, well we’ve got that covered as well. We understand that Data Visualization is a very underrated and a vital technique so we’ve got 9 kinds of Charts such as Bar Chart, Line Chart and Pie Chart to name a few.
Well, you can do that or we have something called a Magic Link for each report which basically anyone can open. Cool !! Right ?
- Model Registry: All the models can be found on one page so that you don’t have to keep looking for a model you trained months back. Here, you can also find the metrics corresponding to each model i.e. Precision, Recall, RMSE, MAE, F2-Score, Silhouette Score, Inertia etc. To make things easier for all the developers around we have introduced features such as exposing the Model as an API and generating Data Drift Report.
- Jupyter Notebooks: Well to all the Developers and Data Scientists, we know that we all have our trust issues and we’d rather write code ourselves. We all love Jupyter Notebooks and know how essential they are. So, we also have a Jupyter Notebook Server which comes along with the Product and it will have access to all the project data. Have a look !
We understand that these features would be a lot to remember when you start working on, hence there are tutorials available in every page.
At FluidAI, we are constantly looking for ways to innovate using AI and Data Science. We also have products such as AI Chatbot and Sanatio (python package for Model Validation), so make sure to check them out too. Please reach out to us at email@example.com if you have more questions, we would be very happy to answer them for you. Thanks for reading and happy learning.