Kristina Halbig | Data Scientist

Hi, I'm Kristina Halbig.

A

Endlessly curious and adaptable data scientist who picks up new tools and techniques with ease, driven to uncover insights from complex data that drive positive change and meaningful impact.

About

Hello! I'm a data scientist with a unique blend of eight years of analytical chemistry experience and a passion for uncovering the stories hidden within data. As a lifelong learner, I'm always eager to expand my skillset and implement new techniques to tackle challenging problems.
My toolbox is well-stocked with Python, Pandas, Scikit-Learn, SQL, JupyterLab, a solid understanding of supervised/unsupervised learning, and natural language processing (NLP). I'm a tenacious problem-solver, driven to deliver high-quality work that makes a positive impact on others.
So, if you're looking for a data scientist who combines technical expertise with a touch of chemistry and a whole lot of determination, you've come to the right place. Let's dive into the data together and unravel its secrets!

Coding: Python, Pandas, Numpy, Scikit-Learn, Data Cleaning, Data Wrangling, API, spaCy, SQL, Keras, Tensorflow
Visualization: Matplotlib, Seaborn, Tableau
Libraries: NumPy, Pandas, OpenCV
Software: Github, JupyterLab, Slack, Streamlit
Machine Learning: Classification, Clustering, Decision Trees, Linear/Logistic Regression, Neural Networks, NLP, Predictive Modeling, Supervised/Unsupervised Learning

I'm seeking opportunities to apply my skills at a dynamic organization that values learning, embraces challenges, and supports career development and growth.

Experience

General Assembly

Data Science Immersive Fellow

Successfully completed a rigorous 12-week data science bootcamp program, totaling over 400 hours of instruction and hands-on learning.
Developed varios open-ended data science projects to explore different issues and improve data skills.
Built, evaluated, and optimized machine learning statistical models and classifications.
Developed strong problem-solving skills through hands-on practice with real-world datasets and case studies.
Tools: Python, Pandas, JupyterLab, Github, Scikit-learn, spaCy, Google Colab

March 2024 - June 2024 | Remote

Encore Labs

Senior Analytical Chemist

Led data wrangling of over 10,000 data points per day, supporting 270 clients and generating $600k in monthly revenue.
Developed and implemented Tableau dashboards to visualize 500 geographic data points, resulting in improved understanding of business performance and strategic decision-making.
Implemented a QBench LIMS system to automate the upload and approval of data, ensuring proper handling of more than 5k data points per day and reducing errors to less than 5% (from 15%).
Trained analysts and lab supervisors on the operation of analytical chemistry equipment and data analysis, ensuring all trainees received scores of at least 4/5 on all KPIs and increasing productivity by 14% within the department.
Tools: Tableau, Microsoft Excel, QBench, Sciex OS (Data Analysis Software)

July 2018 - Nov 2024 | Pasadena, CA

The J.M. Smucker Co

Analytical Chemist

Directed analysis and data wrangling of diverse ingredient datasets across 100 product lines, processing data packets for about 20 samples daily. Led training, oversaw analytical instrument maintenance, and documented lab procedures.
Maintained compliance with industry standard regulations, overseeing the audit of 75 standard operating procedures to ensure compliance and assisting with the acquisition of an ISO 17025 certification.
Performed cost analysis, resulting in $400 daily savings through efficient purchasing.
Tools: Microsoft Excel, Agilent Masshunter (Data Analysis Software)

June 2015 - July 2018 | Burbank, CA

Projects

                
Hack for LA Volunteer
                  Changing LA for the better through data science.
                
AccomplishmentsTools: Data Exploration, Dataset Evaluation
Review datasets from the LA Controller's website, organize them into an Excel sheet with key details, and identify high-value datasets for future initiatives.

Knitting Pattern Recommender
                  A knitting recommender based on Ravelry API.
                
AccomplishmentsTools: Python, Scikit-learn, spaCy, Streamlit, API Scripting
Developed a knitting pattern recommender system using Ravelry API data, implemented in a Streamlit website that suggests similar patterns.

Subreddit Post Classifier
                  A model that categorizes pen-related posts into pens or fountainpens subreddits.
                
                Accomplishments
                Tools: spaCy, API Scripting, Logistic Regression, Naive Bayes 
Scraped posts (titles and content) from Reddit's pens and fountain pens subreddits.  Trained a Naive Bayes classifier to identify the subreddit of each post based on text.

Multiple PDF Reader & Chatbot
                  Chat with multiple PDFs utilizing ChatGPT API.
                
                Accomplishments
                Tools: LangChain, OpenAI API, Streamlit, VS Code
LangChain, OpenAI API, and Streamlit were used to create an app enabling users to upload PDFs and chat about or query their content.

Budget Airline Model
                  Logistic regression model predicting budget airlines from NTSB data.
                
AccomplishmentsTools: HTML, CSS, Bootstrap, Flask, SQLAlchemy, Postgresql, Python
Cleaned and preprocessed NTSB aviation accident dataset with missing and inconsistent data.
Used grid search to optimize model parameters and predict budget airlines.

Home Price Predictions
                  Predictive model for home prices based on various home features.
                
AccomplishmentsEngineered features from a large dataset of home sales, including bathrooms, bedrooms, garage, and square feet.
                    embeddings.
Tested and optimized model performance to avoid overfitting and improve price predictions.