OctWave 3.0 | Workshop 2: Building Models and Handling Time Series Data

Description

The second workshop, titled "Building Models and Handling Time Series Data", was successfully conducted virtually via the Zoom platform on 24th July 2026 from 6.00 pm to 7.30 pm, with the active participation of 45 attendees. The session was led by Mr. Indunil Umayanga, a Data Scientist at PickMe, Sri Lanka. The session commenced sharply at 6.00 pm, with the moderator welcoming all participants. Following this, the moderator introduced the guest speaker, highlighting his professional background and practical industry experience as a Data Scientist at PickMe. This introductory segment concluded by 6.10 pm, setting the stage for an engaging and highly technical session. During the main presentation from 6.10 pm to 7.10 pm , Mr. Umayanga conducted an in depth session on the practical mechanics of building predictive models and managing complex time series datasets which perfectly match to his role as a Data Scientist at PickMe. He began by explaining the unique characteristics of time series data ,such as trends, seasonality, and cyclic patterns followed by essential data preprocessing techniques required to handle time dependent information. He then proceeded to discuss step by step methodologies for feature engineering, model selection, and training. Using real world industry examples, he demonstrated how predictive models are built, optimized, and deployed to solve practical forecasting challenges in fast paced tech environments. The speaker also shared insights into model evaluation techniques specific to time series analysis, encouraging participants to think critically about model accuracy, data integrity, and real-world performance. A Q&A session was held from 7.10 pm to 7.25 pm, during which participants raised insightful questions regarding industry best practices, model optimization, and real time data handling. Mr. Umayanga addressed these queries comprehensively, offering practical industry advice and sharing useful tools and methodologies for aspiring data scientists. At 7.25 pm, the token of appreciation was presented to the guest speaker by the organizing committee, acknowledging his time, effort, and valuable contribution to the success of the workshop. The vote of thanks was delivered, appreciating both the speaker and the 45 participants for their active engagement throughout the session. Finally, the workshop concluded at 7.35 pm.Overall, the workshop was a highly interactive and practical experience that deepened participants' understanding of predictive modeling and equipped them with essential skills for handling time series data in real world applications.