Christopher Hemmens
Verified Expert in Engineering
Data Science Developer
Christopher is a statistician and data scientist with a PhD in behavioral economics. He specializes in mathematical modeling, forecasting, and building tools in Python. He has helped businesses in hospitality, smart home controls, and retirement planning, among others, boost the quality and range of services they offer their clients.
Portfolio
Experience
Availability
Preferred Environment
Python, Pandas, SQL, Data Engineering, Statistical Modeling, Machine Learning, Data Science, Mathematical Modeling, Analytics, Financial Modeling
The most amazing...
...app I've built forecasts retirement savings sustainability decades into the future based on a given strategy of stocks and bonds, which runs in minutes.
Work Experience
Data Scientist
DVX HVAC CO
- Used mathematical modeling to derive the theoretical upper or lower limit for temperature within a space during periods when the heater or cooler was on, generating a target variable for ML models that physically doesn't exist.
- Devised a way to visualize temperature degradation rates to identify location idiosyncracies in terms of rates of temperature change, allowing engineers to tailor solutions to specific locations.
- Devised a way to estimate power consumption for a space, based solely on the square footage, providing the client's customers with a way to estimate future energy costs.
Developer
Second 50 Financial, LLC
- Built a statistical model of the US stocks and bonds markets and used it to generate Monte Carlo simulations for statistically robust retirement savings forecasting.
- Used Markov chain analysis to model the US federal interest rate, providing a realistic baseline above which to forecast individual retirement savings sustainability.
- Built a minimum viable product for forecasting using loops and then solved the problem analytically, decreasing runtime from close to an hour to seconds.
Causal Inference Expert
Jeremy Beasley
- Used panel regression techniques to determine the effect that advertising spend has on revenue, allowing the client to optimize advertising spend.
- Used a difference-in-difference analysis to determine the effect that advertising spend has on revenue, allowing the client to optimize advertising spend.
- Completed full analytical reports on the effects of marketing spend with full data visualization in Jupyter Notebook, allowing the client to easily see key takeaways and business insights.
Scientific Collaborator
HEIG-VD
- Wrote a paper on a novel technique to construct Model-X knockoffs using linear algebra, a framework designed to mitigate the false discovery rate (FDR) in machine learning models.
- Wrote optimization code in various languages, including MATLAB, TensorFlow, and SciPy, and determined the best method for achieving satisfactory code speed and value convergence targets.
- Used variational autoencoders to design index-replicating stock portfolios, thereby mimicking the index fund while minimizing transaction costs.
Data Science Engineer
Comniscient Technologies LLC dba Comlinkdata
- Compiled dozens of reports detailing important differences in two telecoms datasets (big data). Every report included visualizations built in PyPlot and Seaborn designed to highlight key information.
- Conducted complex data analysis on telecoms data from Japan and Canada as a result of unusual activity. Identified key issues and suggested potential causes and fixes.
- Designed reusable dual SQL queries, each one applied to two datasets to be directly compared, one on Google BigQuery and one on Amazon Athena.
- Used geospatial data extensively as a method for aggregating telecoms data (big data).
Data Scientist
Toptal Client
- Built a data cleaning pipeline to produce high-quality data from multiple databases of generic multinational business data as sources, correcting errors, using Google API calls (only) when errors are ambiguous, and identifying duplicates.
- Used PyTorch to build a BERT model to generate a sentiment measure from user-generated reviews of businesses and tourism sites. This measure allowed us to convert scores on a 1 – 5 score to a 1 – 100 score, improving the readability of the reviews.
- Used NLP and Spacy to extract keywords from reviews and identify important features of businesses and tourism sites. This improved tagging and provided new users with prompts for what to leave reviews about, improving site value.
Scientific Collaborator
School of Management, Fribourg
- Conducted extensive research on various trading algorithms with the aid of options data to determine which were the most robust over long-term investment periods.
- Trained a series of machine learning algorithms to see if they outperformed simpler trading algorithms and determined that their improvement was marginal at best.
- Oversaw the work of a PhD candidate and guided him in his research methods and report writing.
Data Scientist
iKentoo
- Built a tool to estimate waiting times in real-time for groups waiting for a table at any restaurant and at any time of day, using statistical modeling and Monte Carlo simulations, improving customer satisfaction for our clients.
- Trained an NLP model on our support tickets to facilitate auto-categorization, allowing our support staff more time to focus on clients' needs and improving efficiency.
- Built a revenue prediction model for restaurants whose prime feature was factoring in weather forecast data. Provided our clients with key information in regard to likely demand.
- Used advanced clustering and data cleaning techniques on our clients' communication preferences to help suggest which of our products would best suit our clients' businesses, significantly improving client satisfaction.
Head of End-user Engagement
Mandat International
- Created publicly available educational resources to promote citizen awareness and knowledge of IoT technology, Smart City interfaces, and the GDPR privacy legislation.
- Led the delivery of reports for the EU's Horizon 2020 program, focusing on end-user engagement.
- Created a series of multi-day educational workshops on smart cities and arranged for them to be broadcast online.
- Conducted research and analysis to compile a comprehensive report on stakeholders within the global cybersecurity ecosystem.
- Wrote and voiced several educational videos detailing the company's projects and outreach efforts.
Journalist
Dukascopy Bank
- Researched and interviewed important individuals from the world of business providing viewers with engaging profiles of them and their work.
- Researched and interviewed important individuals from the worlds of arts, culture, and sport, providing viewers with engaging profiles of them and their work.
- Researched and interviewed important individuals from the world of politics, providing viewers with engaging profiles of them and their work, including the CEO of the Brexit Vote Leave campaign, Matthew Elliott.
Researcher
University of Geneva
- Designed and generated a sentiment measure from financial reporting and performed a time series analysis to determine whether this sentiment was priced in the stock market.
- Conducted causality inference tests to determine whether a particular form of sentiment is priced in the US stock market.
- Built an app in VBA for an academic economics ranking experiment. The app worked in English, French, and German and included well-calibrated randomized elements designed to maximize the information gained from the experiment's dataset.
- Utilized logistic regression models to determine how much of an observed economic outcome can be explained using aesthetic preferences.
- Developed a theoretical model of stochastic decision-making and the endowment effect to help explain pricing differences in two lottery valuation methods. This work won the 2016 SFI Best Doctoral Paper Award.
- Taught Asset Pricing Techniques to master's students at the University of Geneva.
VBA Programmer
Groupe Renault
- Built a tool in Excel macro that forecasts component costs for vehicles based on statistical modeling of currency movements, providing the company with an important risk analysis tool.
- Built a tool in Excel macro with high flexibility in data representation and graph output, which was highly beneficial to the company's market analysts, who could easily see important sources of risk.
- Used high-level statistical modeling techniques to produce highly informative data for our market analysts in regard to risk stemming from forex trading movements.
Researcher
Objective Productions
- Created an outline for a TV pilot about weird facts about human behavior to sell to The Discovery Channel.
- Identified and presented key screen talent as possible presenters for my TV pilot.
- Worked closely with team members to outline the best possible episode topics and episode order for our submission to The Discovery Channel.
Experience
Estimated Waiting Times
Autofill on Support Tickets
Stock Market Sentiment Measure
Smart City Education
Foreign Exchange Projections
Behavioral Finance Research
Review Keyword Identification
Diverse Dataset Merging
Big Data Visualization
Feature Knockoff Construction
Modelling of Temperature Degradation
Causal Inference Analysis
Sustainable Retirement Investing
Trading Algorithm Backtesting
Skills
Languages
Python 3, Python, SQL, Excel VBA, R, JavaScript
Libraries/APIs
Pandas, NumPy, Scikit-learn, SpaCy, Natural Language Toolkit (NLTK), XGBoost, PyTorch, SciPy, TensorFlow
Paradigms
Data Science, ETL, Quantitative Research
Storage
PostgreSQL, Data Pipelines, MySQL, NoSQL, Data Lakes, Google Cloud
Industry Expertise
Trading Strategy Development, Teaching, Cybersecurity, Marketing
Other
Capital Asset Pricing Model (CAPM), Pricing, Pricing Models, Price Analysis, Pricing Strategy, Derivative Pricing, Econometrics, Behavioral Economics, Probability Theory, Stochastic Modeling, Statistics, Statistical Methods, Statistical Analysis, Statistical Modeling, Statistical Data Analysis, Time Series, Time Series Analysis, Machine Learning, Algorithms, Option Pricing, Research, Data Modeling, Data Engineering, Mathematics, Applied Mathematics, Mathematical Finance, Mathematical Modeling, Mathematical Analysis, Data Analysis, Data Analytics, Linear Regression, Analytics, Financial Modeling, Causal Inference, Forecasting, Decision Modeling, Data-driven Decision-making, Models, Modeling, EDA, Exploratory Data Analysis, Monte Carlo Simulations, Regression Modeling, Quantitative Analysis, Backtesting Trading Strategies, Regression, Data Reporting, Minimum Viable Product (MVP), Quantitative Modeling, Financial Analysis, Reporting, Data Scientist, Backtesting, Stochastic Differential Equations, Bayesian Statistics, Statistical Programming, Statistical Learning, Natural Language Processing (NLP), Clustering, Sentiment Analysis, Predictive Analytics, Predictive Modeling, Film & Television, TV Broadcasting, Operations Research, Fintech, Hospitality, Artificial Intelligence (AI), Artificial Neural Networks (ANN), Text Mining, Data Visualization, Physics, User Behavior, Google BigQuery, Stock Trading, GPT, Generative Pre-trained Transformers (GPT), Optimization, Linear Optimization, Language Models, Data Queries, Finance, Deep Learning, Big Data, Neural Networks, APIs, Writing & Editing, Content Writing, Trading, Algorithmic Trading, Clustering Algorithms, Applied Physics, Quantitative Finance, Software Development, Futures & Options, Excel 365, Logistic Regression, Financial Markets, Risk Management, Behavioral Science, Neuroscience, ISO Standards, Smart City Technology, Smart Cities, Educational Videos, Journalism, Internet of Things (IoT), Data Cleaning, Geography, Networking, Financial Data, Variational Autoencoders, BERT, Forex Trading, Forex, Reports
Tools
STATA, BigQuery, Excel 2013, MATLAB, Amazon Athena
Platforms
Jupyter Notebook, Amazon Web Services (AWS), Google Cloud Platform (GCP)
Education
Ph.D. in Finance
University of Geneva - Geneva, Switzerland
Master's Degree in Mathematics
Imperial College - London, England, UK
Certifications
Professional Data Engineer
Google Cloud
JavaScript Basics for Beginners
Udemy
How to Work with Toptal
Toptal matches you directly with global industry experts from our network in hours—not weeks or months.
Share your needs
Choose your talent
Start your risk-free talent trial
Top talent is in high demand.
Start hiring