Citizen Data Scientist Workflow in BDB Platform

by Nikita Agrawal

Figure: 1(a)

Streamline your path to data driven Insights

BDB’s AI-driven architecture seamlessly converts your raw business data into the analytics-ready data sets. It focuses on the entire process from data ingestion to data cleansing and data model creation till visualizing the outcome through single integrated analytics platform.

Data Ingestion is a process of propelling your data to an environment where it is cleansed and transformed, to fit further analysis. Ingest your structured or unstructured data through BDB’s multiple Data Sources like Excel, MySQL, MSSQL, Oracle, and PDF.

Data Cleansing & Transformation - It is a process of removing invalid data and converting it from one format to another. This method improves data consistency and makes data processing easier. With BDB’s Data Cleansing & Transformation tool, perform statistical profiling, identify outliers and detect anomalies using ML.

AI-ML-Predictive Workbench - It is the practice of inferring patterns and predicting future outcomes from existing data sets. BDB’s AI & ML-based workbench helps to accurately predict outcomes that can further be incorporated into the visualization layer.

Visualize & Analyze - BDB provides drag and drop interface to visualize data dimensions and measures quickly. Users can use Text & Voice based responsive Data Search to create fluid visuals on the fly and dive deeper into your data to conceptualize hidden trends by asking critical questions. Gain vital insights into your data to increase productivity even when on the go.

Figure: 1(b)

Data Ingestion, Cleansing & Transformations

Ingest structured, semi-structured or unstructured data from various data sources. Perform statistical profiling, remove non-essential data, fix inconsistencies, apply several transformation techniques to analyze and recognize your data set.

  1. Data Profiling / Sampling
  2. Data Cleansing / Quality
  3. Data Integration
  1. Data Reduction / Modification
  2. Data Transformation

AI/ML Workbench

A Workbench with Integrated Algorithms from R, Spark ML, Python, Keras + Tensorflow for Citizen Data Scientist to create Algorithmic workflows, Prescriptive flows to accurately predict the future business outcomes. Users can collate data through data pipelines, transform feature sets, train/test, compare and deploy ML models into ML pipelines, to be consumed as an ML service within our self-service BI dashboards.

Experience the most intuitive and interactive predictive models through our web-based UI that brings together advanced analytics, optimization, real-time scoring, and machine learning to analyze the ‘HOW’ and ‘Why’ factors behind each business move and helps to decide the next probable business strategy.

  1. Visualize market trends & Take actions
  2. Custom Scripts
  3. Classification Techniques
  4. Predictive Modelling
  5. Relational Insights

Keras + Tensorflow Workflow

Self Service BI capabilities

  1. Drag & Drop Interface
  2. Customize chart types and Properties
  3. Web socket Integration
  4. Custom Calculated Measures
  5. Integration with R, NN
  1. Collaboration Features
  2. Robust Filter Panel
  3. Explore Millions of Rows in a Flash
  4. Intuitive UX & Data Search
  5. Export Options & Alerts

AI based Report

Self Service enhanced by AI based Search

Data Discovery
Access a user-oriented process, for detecting patterns and actionable insights. The goal is to make messy and scattered data clean and comprehensible.

Data Inquiry
Examine different slices of data and hone their inquiry while inspecting data. This process generates a sequence of queries on the underlying data sets.

Explore your transformed business data visually to identify patterns & relationships through BDB’s rich charting library.

Since the data set can be huge in volume, there needs to be a mechanism to get the pertinent and significant information. Data summarization techniques aim to produce summaries, saving time and resources. Techniques like Context Sensitive Aggregate Engine, clustering, bucketing for histograms, linear regression etc. are used.

Conversational Analytics on Mobile

Our NLP based AI driven voice assistance lets you drill deeper into your data and create a complete visual report without even using your fingers. Users can even customize the search to reflect jargon of their organization via adding 'Synonyms' to the data search library. Intuitive search Autocompletes the user query while the user types. Data can be filtered up to desired level of granularity. Now, typing a search query is supported in Mobile.

BDB empowers users in making rational business decisions by giving them deep business insights in seconds. Access your data anytime and anywhere from your system or mobile (Android or iOS) by asking queries via text or speech.