Top 10 DataHours – An Overview
Analytics Vidhya is one of India’s largest communities for Data Science, Machine Learning, and Analytics. We have always kept our community at the centre stage of the ecosystem. From blogs and practice problems to courses and hackathons, our community has been actively engaged with us throughout the years to share and disseminate knowledge in the data science domain.
Keeping the community in mind, at the beginning of 2022, Analytics Vidhya initiated a one-hour webinar series called: “DataHour’.” We have completed over 170 sessions, and more than 17,000 attendees have leveraged the sessions. The top industry leaders conduct these sessions, covering domains ranging from AI, Blockchain, Career-oriented guidance, Data Science, Data Visualization, ML, NFT, NLP, etc.
While we are on the edge of entering a new year, it’s about time that we reflect on the top 10 DataHours of 2022 based on the popularity of the Datahour among viewers.
Artificial Intelligence in Retail
Retail is the art of selling things. Have you ever thought about how ML and AI can revolutionize traditional retail?
In this DataHour Dr. Shantha Mohan, (Mentor and Project Guide at Carnegie Mellon University, Integrated Innovation Institute) vividly explained how Artificial Intelligence and Machine Learning could create and enable entirely new ways of doing business—starting from the evolution of the retail industry, its current landscape, future potential, products, and channels. She covered every aspect of the retail industry and demonstrated an E-commerce use case of AI/ML in the retail industry. If you want to master the skill of efficiently doing business in the current state, then this DataHour is for you! To know more, Click on LINK.
Duration: 1:11:29
Key Points Covered: History, current situation, and future image of the Retail Industry, AI/ML in E-commerce, and Supply Chain Management.
Hands-on With Data Analysis Expressions (DAX) for Power BI
Power BI is one of the most preferred tools for Data Visualization, and one can never learn Power BI without DAX. This DataHour inculcated the skill of DAX along with tips and tricks for the viewers. The methodology of KPI, row and filter context, calculated column and calculated function, the most interesting Waterfall chart, and new and quick measures – all of it was covered in this DataHour by Pranav Dar (BI & Analytics – Trainer & Consultant). To know more, Click on LINK.
Duration: 1:06:53
Key Points Covered: Calculated Function, Waterfall Chart, and New measures in DAX.
Create Effective Data Science Notebooks and Communication
You will agree with me when I say that communicating the insights and codes is as important as finding and running them, right? An Ideal Notebook should provide accurate insights and express the story behind the data clearly. In this DataHour, Martin (Data Scientist at YipitData | Kaggle Grandmaster) recommended a few tips on designing impactful notebooks according to your needs for effective Data Science communication. He explained the exploratory notebooks, data validation, data plotting from different angles, and effective ways of documenting data. To know more, Click on LINK.
Duration: 1:01:38
Key Points Covered: Notebook maintenance, Data plotting, and validation.
Introduction to Interpretable Machine Learning
The need for interpretable Machine Learning arises because of the incompleteness of machine learning as a solution, and this DataHour by Serg Masís(Climate & Agronomic Data Scientist at Syngenta) perfectly fulfills the need. It explains why, how, and when model interpretation is done while covering the use of scikit-learn library, Ag boost, and Cat Boost. The process of error breakdown, logical regression, decision trees, correlation coefficient, and networks were also discussed. To know more, Click on LINK.
Duration: 1:09:56
Key Points Covered: Machine learning model, correlation, regression, and decision trees.
Traversing the Journey of an Analytics Problem
As with everything in the universe, Data Analysis also flows through a journey. To learn how to decode the real-world problem using analysis and find meaningful results, this DataHour by Amitayu Roy (Senior Manager (Applied Intelligence, Strategy & Consulting) at Accenture) was the best stop. This DataHour explored the various aspects of data, data engineering, and the data science industry. It also traced the problem journey, customer journey, chun retention, hypothesis-driven approaches to solving an analytics problem, and ways of deriving useful insights. To know more, Click on LINK.
Duration: 1:30:20
Key Points Covered: Data Science, Data Engineering, Problem journey, Churn Retention
Introduction to Image Processing using Python
This DataHour by Siddhant Sadangi (Data Scientist at Reuters) was a code-along session that taught the digital representation of images and how image characteristics can be altered using python. This DataHour was centered around image analysis, Greyscale imaging, Color channels, ways of twerking image properties, and identification of duplicates.To know more, Click on LINK.
Duration: 59:55
Key points covered: Greyscale imaging, Color channels, Image properties alteration.
Building NLP Applications using Hugging Face
Hugging face is the savior of developers as it provides tools that make the addition of state-of-the-art transformer models into the applications easier. Through this DataHour, Julien Simon (Chief Evangelist at Hugging Face) explained to the audience about the transformer models, how we could solve business problems using them, and the process of accelerating machine learning projects end to end. He also differentiated between transfer learning and pretrained models and discussed the working of Pipeline API, hardware accessories, tokenized training set, model card, and accuracy targets. To know more, Click on LINK.
Duration: 1:01:14
Key Points Covered: Pipeline API, Pretrained Models, Tokenization of training set, Model Card.
Data Visualization using Python
The power of visualization is evident in all aspects, and the data domain is no exception. This DataHour by Nitish Vig (Business Analyst at Trejhara Solutions Ltd) covered how we can visualize data using python, construct charts, and relate them with other applications like excel and Power BI. The Speaker elaborated on the use of visualization tools, methodology of univariate and bivariate analysis, the difference between numerical vs. categorical data visualization, comparison between Googlecolab and Jupyter, and charts including count plot, scatter plot, line charts, strip plot, heat maps, etc. To know more, Click on LINK.
Duration: 1:17:53
Key Points Covered: Visualization tools, Charts, Univariate, and Bivariate analysis.
Introduction to MLops
Machine Learning models follow complex life cycles, from exploration, deployment, and maintenance. MLops is a set of practices for deploying and maintaining ML models reliably and efficiently.
The speaker of the DataHour, Anmol Krishan Sachdeva (Hybrid Cloud Architect at Google), descriptively explained MLOps, covering model operationalization using the power of Airflow and Kubernetes. This DataHour briefed about model registering, data acquisition and pipeline, data erroring, model metadata, goals and tools of MLOPs, and apache airflow. To know more, Click on LINK.
Duration: 1:02:06
Key points covered: MLOPs, Data Acquisition, Data Erroring, and Tools of MLOPs.
Data Science in a Faang Company
You probably would have heard about FAANG if you didn’t live under a rock. FAANG is an acronym of Facebook (now Meta), Amazon, Apple, Netflix, and Google (now Alphabet). We all have witnessed the impact of Facebook in the world of technology and human beings. In this DataHour, Eshan Tiwari (Data Science Lead at Facebook) walked us through how data science is being used in Facebook and how one can become a data scientist at Facebook. He guided the audience about the role of data and data science in FAANG, and provided examples of data science projects performed in FAANG. To know more, Click on LINK.
Duration: 54:37
Key points covered: Data science use in FAANG, career in FAANG in the data science domain.
Conclusion
Though we have listed the top 10 DataHours, each DataHour was unique and interesting in its way, with diverse topics and expert speakers. Through these DataHour series, we provided our viewers easy access to the latest and trending topics in Data Science and technology domain. In the next year, we will continue to deliver the same with increased frequency and topics. So let’s wrap up this year with these top 10 DataHours and wait for the next list!
