Data Science Philadelphia (DataPhilly)

Philadelphia, PA, US

5,979 members · Public group

Organized by Michael Becker and 5 others

What we’re about

DataPhilly is a community run group for anyone interested in gaining insights from data. Topics include (but are not limited to) predictive analytics, applied machine learning, big data, data warehousing and data science. We <3 data!
Join our slack to help plan future events! http://bit.ly/DataPhillySlack
Email us at: admin[at]dataphilly.com
Found a space we can use for future meetups? https://goo.gl/Ru0eth
Found a speaker for an upcoming meetup? https://goo.gl/9DJxq0
Found a sponsor for our events? https://goo.gl/JLVfqh

DataPhilly's Code of Conduct

Want to have access to the video recordings and details of our past events? https://dataphilly.github.io

Upcoming events (4+)

See all

Thu, Jan 23, 2025, 11:30 PM UTCData Philly Study Group: Session 1
Link visible for attendees
Welcome to the first 2025 DataPhilly Study group session, where we will work on Kaggle competition
Kaggle competition to follow will be announced on Jan 4 2025.

Week 1 Activities:

Kaggle Setup:

Create a Kaggle account if not already done.

Explore the Kaggle platform, datasets, notebooks, and leaderboard.

Set up a GitHub repository to track project progress.

Competition Overview:

Study the competition's problem statement and evaluation metric.

Understand the data: Download the dataset and explore the files (e.g., training set, test set, sample submission).

Discuss the data dictionary and problem requirements in the context of the dataset.

Data Exploration (Exploratory Data Analysis - EDA):

Perform basic data exploration (checking data types, null values, etc.).

Visualize the data (e.g., histograms, box plots, scatter plots) to get insights into data distributions and relationships.

Assignment:

Write a blog or GitHub post explaining the competition’s problem and data.

Share initial findings from EDA.

We will add meetup link soon.
49 attendees+44
Thu, Jan 30, 2025, 6:00 PM ESTDataPhilly Talks: Leadership in the AI Era and Compelling storytelling with data
BrainDo, Philadelphia, PA
📢 Join us for the first DataPhilly Talk of the year in collaboration with BrainDo and Philly Analytics!
🤖 Learn about Leadership in the (Agentic) AI Era and how to craft compelling stories with data.
Our host, BrainDo, is a Philadelphia-based team of digital marketing experts who partner with clients to create meaningful experiences with data-driven results. They also host and manage Philly Analytics, with monthly Analytics Community Meetups called Web Analytics Wednesdays.
More details at phillyanalytics.org or the Web Analytics Wednesday group on LinkedIn.

Speakers:
Andreas Welsch
Founder & Chief AI Strategist, Intelligence Briefing
Author of the AI Leadership Handbook
https://intelligence-briefing.com
Talk: Between Automation and Autonomy: Leadership in the (Agentic) AI Era

Summary:
The pressure on leaders to incorporate AI into their business has never been higher. Despite their best efforts, 85% of AI projects fail. But that is rarely just because of technology. Between unrealistic expectations of what AI can do, isolated science projects turned into money pits, and employees looking for clarity on how AI influences their jobs, learning how to successfully introduce AI into your organization can be difficult. Getting it right on the second or third attempt is not an option. In order for your AI program to be successful, you need to be aware of all the facets involved, not just the technology—a daunting task. Spanning strategy, stakeholder management, collaboration, culture, ethics, data privacy, risk management, and technology, Andreas covers everything you need to know to become a confident and successful AI leader—and get it right on the first try.

Abstract:
Agentic AI is set to take center stage in the industry in 2025 with AI labs building the foundational capabilities and software vendors incorporating it quickly into their platforms and applications. But what does it mean when AI takes on more complex tasks? How will this adoption happen and how can technologists lead this new way of working?
Just like during the previous Machine Learning and Generative AI hypes, business leaders will be compelled to find quick wins for this new technology. But in their rush to adopt Agentic AI, they will miss critical steps and aspects (again): Aligning AI with their business strategy, managing change, and taking a responsible approach to AI.
Between hype and hope lies the potential of Agentic AI to drive meaningful business impact. From single agents to departmental multi-agent systems, inter-departmental and inter-company agent systems, the progression towards increased autonomy and automation is already apparent. Data professionals and leaders play a critical, yet often underestimated role in providing the data required for these new systems to perform. Together with their business stakeholders, they can ensure that the outcomes and results that Generative AI and Agentic AI systems generate are timely, accurate, and valuable.

Speaker Bio:
Andreas Welsch is an internationally recognized AI leader, advisor, and speaker known for helping the world's leading enterprises turn technology hype into business outcomes. He is the Founder & Chief AI Strategist at Intelligence Briefing, the best-selling author of the AI Leadership Handbook, and the creator of several popular LinkedIn Learning courses on AI agents for business leaders.

Dasanj Aberdeen
Product Management and Content Leader
Talk: From Numbers to Narrative: Crafting Compelling Stories with Data

Summary:
Data is only as powerful as the story it tells. In this session, we’ll discuss how to transform raw numbers into narratives that inspire action, influence stakeholders, and communicate insights effectively. You’ll learn techniques for identifying key takeaways, crafting a narrative arc, and visualizing data to enhance clarity and engagement. Whether you're pitching a campaign, delivering a performance update, or influencing product strategy, this talk will give you the skills to make your data insights memorable. Attendees will engage in hands-on exercises to practice storytelling techniques and leave with tools to elevate their data presentations.

Abstract:
In the age of information overload, it’s not enough to present data - you need to tell a story. This presentation will teach you how to transform raw data into compelling narratives that engage audiences, drive decisions, and inspire action. We'll explore the three pillars of data storytelling: identifying the key message, building a narrative arc, and visualizing data effectively. Using relatable examples and hands-on exercises, you’ll learn how to simplify complex data into stories that resonate with diverse audiences.
The interactive session will include a hands-on activity where attendees collaborate to create a data-driven story using a real-world scenario. You’ll gain practical skills for using frameworks like Problem-Insight-Action and learn how to tailor your message to stakeholders at different levels of the organization.
Whether you're persuading executives, pitching creative campaigns, or analyzing performance metrics, this talk will equip you with the tools to make your data presentations unforgettable. Join us to move beyond the numbers and unlock the power of storytelling in your work.

Speaker Bio:
Dasanj Aberdeen is a Product Management and Content Leader, creator, and champion of interdisciplinary pursuits. She brings digital transformation expertise in AI product management, agile project management, digital marketing, and data across Fortune 100 companies in technology, media, and telecommunications, financial services, and healthcare industries. Dasanj is an Adjunct Professor, empowering the next generation of well-rounded, business-savvy engineering graduate students. She is a graduate of the Innovation Management and Entrepreneurship program at Temple University’s Fox School of Business and the Wharton School at the University of Pennsylvania.

🎊 We look forward to seeing you there!
39 attendees+34
Thu, Feb 6, 2025, 11:30 PM UTCData Philly Study Group: Session 2
Link visible for attendees
Week 2 Activities:

Data Cleaning:

Handle missing values (imputation, deletion, or other strategies).

Address any outliers or noisy data.

Convert categorical features to numeric (e.g., using one-hot encoding, label encoding).

Feature Engineering:

Create new features based on the data (e.g., interaction terms, domain-specific features).

Normalize/standardize features if required (for certain algorithms like SVM or neural networks).

Split the data into training and validation sets.

Modeling Introduction:

Introduce simple models: Linear Regression, Decision Trees, or Random Forest.

Train these models and evaluate their performance using cross-validation.

Assignment:

Share cleaned and preprocessed data, along with initial features on GitHub.

Write a blog post explaining data preprocessing steps and feature engineering techniques used.

We will add meetup link soon.
12 attendees+7
Thu, Feb 13, 2025, 11:30 PM UTCData Philly Study Group: Session 3
Link visible for attendees
Week 3 Activities:

Model Selection:

Explore different models such as Logistic Regression, Gradient Boosting (e.g., XGBoost, LightGBM), and Neural Networks (if applicable).

Compare model performances on the validation set.

Use a scoring metric from the competition to evaluate the models (e.g., accuracy, AUC, F1-score).

Hyperparameter Tuning:

Tune hyperparameters using grid search or random search.

Introduce concepts of cross-validation and overfitting.

Implement early stopping and regularization techniques to avoid overfitting.

Ensemble Methods:

Introduce model ensembling techniques such as bagging, boosting, or stacking.

Combine predictions from different models to improve accuracy.

Assignment:

Share trained models, hyperparameters, and results on GitHub.

Write a blog post explaining how to select and tune models and why ensemble methods work.

We will add meetup link soon.
10 attendees+5