1 Day Workshops

Focus on a key topic for a day

DSFM offers one-day workshops to focus on a particular topic of interest. Perhaps you've taken the DSFM Boot Camp or Fast Track and you need to explore new models, concepts, or options to address a particular problem; or perhaps you've taken the DSFM Boot Camp or Fast Track and now you wish to return with your Data Science team to work on a particular problem; or perhaps you are new to DSFM and one-day Workshops are a good way to try out and get started in the program.

Each Workshop focuses on just one particular issue so you have time to learn a broader range of tools and concepts to work with in that area - no need to spend time in areas you may not need. Workshops are also easier to fit into a busy work schedule without having to plan for extended time away from the office.

      • Each Workshop runs from 10:00 AM to 3:00 PM.

      • The short-day allows you to commute to Lausanne, and/or stay in touch with colleagues at work.

      • We provide lunch - but it is a working lunch so that we have more time for discussion.

Please see the enrollment page for workshop dates and options.

Python Quick Start

Most courses in the DSFM program use Python for programming demonstrations, exercises, and projects (when applicable). Python is the most popular, and widely-used, programming language for Data Science. This Workshop provides an accelerated introduction to the syntax, objects and packages of Python that are most important for processing data, analyzing data, and building predictive models. Most of the complicated programming of Data Science is encapsulated into dedicated packages (such as Pandas or Scikit-Learn), so even an introductory, basic familiarity with Python can help you be much more productive.

Topics Covered:
Workshop covers the syntax, data structures, packages, and logical flow used in a typical Python Data Science Project. Because 80% of a Data Science project is often spent on data pre-processing, cleaning, inspection, visualization, and summarization - basic Python skills can be very important to the project.

Preparation:
No preparation is required to take the course - but a prior understanding of the basic principles of computer programming is strongly recommended. We therefore recommend covering the introductory topics of the JetBrains Academy Python Track prior to the workshop - you will then arrive and at least know how to get started. You do not, however, need to take the entire JetBrains Python Track - completing the JetBrains Python Track can take up to 30 hours to complete and will duplicate much of the material from the Workshop.

Dates:
See the enrollment page for the next time this workshop is offered.

Text As Data

Text is an increasingly important aspect of business analytics and making business decisions. Text, however, is often captured in non-structured formats, and analyzing text requires specialized programming tools and machine learning models.

Topics Covered:
Workshop covers the entire Text Analysis Pipeline, including: natural language pre-processing, text embedding, topic modeling, sentiment analysis, and outcome prediction. The workshop focuses on methods that work on textual big data, and methods that can run at scale. Methods covered include: text pre-processing, spaCy, TextHero, TF-IDF, Word2Vec, Glove, FastText, and Attention-based methods for text. Models covered include Latent topic modeling (LDA, etc.), Recurrent Neural Networks (RNN, etc.), and Deep bidirectional transformers (BERT, etc.)

Preparation:
This Workshop requires basic to intermediate knowledge of Python programming - at least at the level you would have after completing a DSFM Boot Camp course. We therefore recommend taking a Python Quick Start Workshop, the DSFM Boot Camp, or the JetBrains Academy Python Track before the start of this Workshop.

Dates:
See the enrollment page for the next time this workshop is offered.

Model Evaluation

So you've built a model - but is it any good? Understanding why a model makes a particular prediction is central to building an effective data science application. Does your model meet your objectives? Does it perform better than common sense? Are outcomes economically meaningful? Are predictions biased or harmful to particular subpopulations of the data? Increasingly complex models, however, can turn them into a “black box”, making it hard to interpret model behavior and erode trust among stakeholders.

Topics Covered:
Workshop focuses on learning how to make complex models more interpretable, so that their results can be communicated better to others and acted on more effectively. Linear regression and tree-based models are already widely used precisely because they are easier to interpret. But more advanced, machine learning and neural network models can often outperform the older, traditional models. This workshop covers on model-agnostic methods - such as permutation feature importance and Shapley values - to give analysts a better understanding of what complex models do, and why they make the predictions that they do. The Workshop will go through the core topics conceptually, our emphasis will be on equipping participants with a range of practical computational tools to take back and use in their workplace.

Preparation:
This Workshop requires basic knowledge of Python programming - at the level you would have after completing a DSFM Boot Camp course. We therefore recommend taking a Python Quick Start Workshop, the DSFM Boot Camp, or the JetBrains Academy Python Track before the start of this Workshop.

Dates:
See the enrollment page for the next time this workshop is offered.

Fraud Analytics

As more and more business goes online, the detection of fraudulent activity has become an increasingly important topic for many businesses. The consequences of fraud can be very expensive - for the bottom-line, for a firm's reputation, and for customer satisfaction if measures to avert fraud are too aggressive. This Workshop introduces you to these tradeoffs, and the machine leaning methods and models that help you to optimize your outcomes.

Topics Covered:
Workshop covers the conceptual issues of identifying fraud, examples of fraud, traditional rule-based-approaches to detecting fraud, and finally updated machine learning methods for outlier/anomaly detection. Because fraud is typically rare, the Workshop also covers the analysis of unbalanced datasets. Fighting fraud can also involve new sources of data from Social Networks, so the Workshop also provides an introduction to networks and fraud.

Preparation:
Most of this Workshop is conceptual; working demonstrations and examples, however, will also be provided in both Python and R. To get the most out of the practical materials, you may want to complete the basic Python course at Kaggle (~ 7 hours if you are new to Python) or take the Python Quick Start workshop before attending this workshop.

Dates:
See the enrollment page for the next time this workshop is offered.

Corporate Integration

Developing and deploying Data Science models in large, corporate environment scan pose unique challenges and opportunities. Data is often held in specialized databases, Enterprise Resource Planning (ERP) systems, and accounting systems. Corporate IT platforms provide unique tools such as Power BI, Tableau, etc. Large amounts of valuable data are still hidden away in Excel spreadsheets, CSV files, and other formats. All of these conditions may require manual pre-processing, cleaning, merging, and other steps to access the data, analyze, visualize, and eventually put into use.

Topics Covered:
This workshop focuses on Python as a tool that can be used across the corporate landscape to integrate between different systems and break-down information silos. It will focus on exploratory data analysis, the steps needed to bring data into a dashboard, the application of Python with other tools (pure Python vs. Excel-PowerBI-Tableau) and python-script integration across these tools, and good practices and standard rules for visualizing data.

Preparation:
This Workshop requires basic knowledge of Python programming - at the level you would have after completing a DSFM Boot Camp course. We therefore recommend taking a Python Quick Start Workshop, the DSFM Boot Camp, or the JetBrains Academy Python Track before the start of this Workshop.

Dates:
See the enrollment page for the next time this workshop is offered.

Forecasting & Monte Carlo Simulation

Forecasting and simulating outcomes are two of the most requested (and yet under-represented) aspects of Data Science. Vision recognition, language translation, and chatbots grab the most attention, but simply forecasting activity or sales going into the future - and simulating high/low/probable outcomes - often are the most important problems faced by real companies. This workshop addresses that gap and aims to teach you real, applied models that you can take back to the forecasting and simulation problems you face every day.

Topics Covered:
For time-series forecasting, Workshop covers traditional regression models (ARIMA, Seasonal ARIMA, etc.), non-parametric and locally-linear models (lowess, locally-linear, polynomial models, etc.), deep neural network models (Prophet, GluonTS, etc.), and high-level utilities (Dart, etc.). For Monte Carlo Simulation, the Workshop covers Crystal Ball in Microsoft Excel, pure-Python implementations, and basic Markov Chain Monte Carlo methods.

Preparation:
This Workshop requires basic to intermediate knowledge of Python programming - at least at the level you would have after completing a DSFM Boot Camp course. We therefore recommend taking a Python Quick Start Workshop, the DSFM Boot Camp, or the JetBrains Academy Python Track before the start of this Workshop.

Dates:
See the enrollment page for the next time this workshop is offered.

Cloud Computing

Big Data often requires using No-SQL databases, virtual machines, remote access, UNIX commands, Spark clusters, transportable containers, Kubernetes orchestration, and platform-as-a-service (PaaS) solutions running in the cloud. As such, cloud computing is quickly becoming an essential part of the machine learning and data science landscape. Cloud computing, however, also requires an entirely new set of system administration and user-operator skills, as well as a new conceptual understanding of what-does-what, and how it all fits together. Cloud computing can be fast, powerful, and scalable - but cloud computing can also be bewildering to the newcomer.

Topics Covered:
Workshop focuses on system administration and practical use, with real examples running on Google Compute, Amazon AWS, and Microsoft Azure. Basic knowledge of Python and UNIX commands are helpful, but not essential. Please note that the course focuses on the practical aspects of running individual projects on the cloud - and not enterprises-level aspects of setting up or designing cloud architectures, data lakes, or other large IT system/ERP system integration. This course is for the individual data scientist (and/or their team) - not a broad IT department.

Preparation:
The Cloud Computing Workshop is a hands on course with practical examples running as UNIX bash scripts, Python programs, Spark-Python programs, and other system administration tasks. You do not need to have advanced skills in any of these domains, although some familiarity with Python, bash, and SQL will help you get more from the course that you can immediately use. We therefore recommend taking a Python Quick Start Workshop, the DSFM Boot Camp, or completing the basic Python course at Kaggle (~ 7 hours if you are new to Python).

Dates:
See the enrollment page for the next time this workshop is offered.

Data-Driven Strategy

The Machine Learning Revolution in algorithms, the 4th Industrial Revolution in cyber-physical automation, and the Digital Transformation of many companies, are all changing how strategy impacts the firms. Senior executives now have to plan for, and make decisions about, a broad array of technologies and conditions for which they have little formal training. Nevertheless, strategic decisions today are as important as ever.

Topics Covered:
Workshop focuses on the economic fundamentals of strategy, and how those principles apply to the digital and algorithmic world. The importance of algorithmic prediction vs. managerial judgment has shifted, the role of automation and business robotics are challenging old operations, evidence-based reasoning is challenging the Highest Paid Person's Opinion (HiPPO), and competitors are sprinting forward with completely new business models. Against this background, senior management needs a refresh and update as to what strategy means in the digital age.

Preparation:
This course involves zero programming - no preparation needed.

Dates:
See the enrollment page for the next time this workshop is offered.

Litigation Support

Attorneys often faced difficult decisions to be made under high degrees of uncertainty. This course adopts new methods of analysis from data science, computational thinking, and Bayesian updating to guide practitioners to better decisions. All of these topics are becoming increasingly important to the practice of law.

Topics Covered:
DSFM is still developing this workshop, but we intend to focus on methods to operationalize decision-analytics for litigation strategy. The course combines theoretical models, simulated scenarios, and empirical data. More details coming soon.

Preparation:
More details coming soon.

Dates:
This course is still in development.


Next Steps

Please see the enrollment page for course dates and options.

Sign up for our newsletter to receive updates about the program.

Or contact us if you have questions.