Fernando Melchor, Developer in Mexico City, Mexico
Fernando is available for hire
Hire Fernando

Fernando Melchor

Verified Expert  in Engineering

Data Scientist and Software Developer

Mexico City, Mexico

Toptal member since July 3, 2021

Bio

Fernando is a data scientist with a background in mechatronics engineering. He is passionate about cities and all the data that describes them, especially geospatial data. During his career, he designed, prototyped, and launched data products used worldwide in diverse sectors, such as automotive, government, on-demand streaming, and real estate. He has experience deploying machine learning models, data science projects, ETL processes in Airflow, and KPI initiatives in Tableau and Metabase.

Portfolio

Flat.mx
Python, PostgreSQL, Dash, Plotly, Scikit-Learn, TensorFlow, Docker...
Mexico City Government
Python, Pandas, SQL, Python, Plotly, Spark, SVMs, Scikit-Learn, Neo4j, Python...
Discovery, Inc.
Pandas, Python, SQL, Hadoop, Python, Plotly, Dash, Tableau Development, Cron...

Experience

Availability

Part-time

Preferred Environment

Python 3, Apache Airflow, Amazon Web Services (AWS), Tableau, GIS, Spatial Analysis, FastAPI, Machine Learning, Deep Learning, Data Visualization

The most amazing...

...thing I've done was create and execute a data strategy and vision for a startup company.

Work Experience

Chief Data Scientist

2020 - PRESENT
Flat.mx
  • Developed machine learning predictive models for real estate properties in Mexico City, including long- and short-term rent and selling prices.
  • Built a machine learning DevOps framework to easily deploy and update models on Amazon Elastic Container Service (Amazon ECS).
  • Developed an automated offer system that reduced visits to offer lead time from 13 days to minutes.
  • Created data visualization dashboards to democratize data and insights inside the company. Led the KPI efforts to measure every business unit with continuous QA and improvement activities.
  • Developed complex Airflow pipelines to curate data from multiple sources and formats to feed the acquisition team with high conversion rates, leading to scaling the business.
  • Modeled the public transportation network of Mexico City to understand the access and centrality of different areas.
  • Deployed and developed multiple image recognition, classification, and similarity algorithms using deep learning frameworks (TensorFlow, PyTorch).
  • Created and developed a geographical framework and tools to enrich data based on spatial data such as address homologation, street noise index, and access to the Green Space Index.
Technologies: Python, PostgreSQL, Dash, Plotly, Scikit-Learn, TensorFlow, Docker, Amazon Elastic Container Service (ECS), GitHub, Python, Visualization, Tableau Development, GIS, Sedona, GIS, Spatial Analysis

Director of Data Architecture and Data Analysis

2019 - 2020
Mexico City Government
  • Developed the city’s security dashboard for the city’s police department for everyday reporting and crime tracking. This dashboard is a tool that is used daily for crime tracking and decision-making.
  • Created optimization algorithms for police distribution on the subway system, including a visualization tool with metrics and a graphic scheduler.
  • Diagnosed emergency response time of ambulances identifying the three main root causes of delays. The actions implemented reduced ten minutes the average response time.
  • Mentored a team of five data analysts and scientists with best practices and product development methodologies. Implemented on-hands training to develop ETL pipeline and database capabilities on the team.
  • Created a data analytics team to assist the mayor and other city departments with decision-making.
  • Presented and produced multiple exploratory data analyses to inform decision-makers regarding security, emergency response, and mobility.
Technologies: Python, Pandas, SQL, Python, Plotly, Spark, SVMs, Scikit-Learn, Neo4j, Python, Visualization, Tableau Development

Data Scientist | Researcher | Digital Analytics and Insights

2018 - 2019
Discovery, Inc.
  • Developed a data-product dashboard (Tableau) based on app reviews, including an automated ETL pipeline that gets the new reviews and ratings. At the core, I implemented a review text classification model using SVM, achieving 80% accuracy.
  • Developed the alarm dashboard in Tableau that helped to track performance across different platforms and the discovery ecosystem. It allowed exploring the daily performance of eight variables with an anomaly detection system.
  • Created a bipartite graph of streamers and shows for audience clustering. Performed network analysis to understand how audiences overlap in each channel. This data product is used by marketing and programming to guide their business strategies.
  • Implemented a network analysis framework to analyze a market research survey creating a visualization tool for the team to explore the survey results adding demographics filters, helping them design future products.
  • Developed the ROI marketing campaigns dashboard (Tableau) to measure marketing campaign performance. The dashboard helps visually understand the campaign lifecycle and forecasts the expected ROI.
Technologies: Pandas, Python, SQL, Hadoop, Python, Plotly, Dash, Tableau Development, Cron, NLP, Generative Pre-trained Transformers (GPT), Gensim, SpaCy, Support Vector Machines (SVM), Scikit-Learn, Python, Visualization

Data Scientist

2017 - 2018
ARGO Labs, California Data Collaborative
  • Created an ETL pipeline that combines data from water utilities, a web scraper, and public APIs to identify the business type of a water user (commercial or institutional).
  • Designed a classification method that uses data from APIs and NLP to assign a business type to each customer.
  • Created an automated process to aggregate the census data from the block-group level to the water districts level. This allowed the water utilities to understand their customers and the research team to create an analysis using demographics.
  • Led a team of four to develop a water usage benchmark in CA by using publicly available data and water utility data.
Technologies: Dash, Python, Generative Pre-trained Transformers (GPT), NLP, Statistics, Benchmarking, Python, Visualization

MTA Subway Network and the L Train Closure Impact

https://github.com/fernandomelchor/L-Train_Project
Led the strategy, identification, and characterization of the affected areas to design a better contingency plan, measure the impact, and simulate subway users' behavior in case of a disruption using census data and spatial analysis.

Simulation:
https://github.com/fernandomelchor/NYC_MTA_Subway_Network

Automated Classification of Airbnb Listings in NYC

https://github.com/fernandomelchor/Airbnb_Project/blob/master/Airbnb_Paper.pdf
Developed an innovative Airbnb classification method based on cost, transportation connectivity, and businesses in the area.

• Developed a geospatial scan algorithm that gathered characteristics of the area surrounding the Airbnb listing.
• Created a transportation connectivity index based on the subway network using graph theory and spatial analysis.
• Implemented clustering techniques and created visualization maps.

Analysis of the Mexican Senate — Published by Nexos

By web scraping, I created a database of the senators and their votes. I applied PCA to visualize the senators' distribution identified by name and party. Then, I compared each senator to each party to uncover real tendencies beyond the official parliament groups. The data helped me to create a compelling story about the parliament and intra-party dynamics.
2016 - 2017

Master of Science Degree in Data Science and Urban Informatics

New York University - New York, USA

2006 - 2010

Bachelor's Degree in Mechatronics Engineering

Tecnológico de Monterrey - Mexico City, Mexico

MAY 2020 - PRESENT

Data Science For All: Latin America 2020

Correlation One and SoftBank Group

Libraries/APIs

Python, Pandas, Python, Scikit-Learn, TensorFlow, REST API, SpaCy

Tools

Tableau Development, Plotly, Apache Airflow, Cron, Gensim, Amazon Elastic Container Service (ECS), GitHub, GIS

Languages

Python, SQL, Python, JavaScript

Paradigms

ETL

Platforms

Jupyter Notebook, AWS, Docker

Storage

PostgreSQL, Neo4j, Hadoop, GIS

Frameworks

Spark, Sedona

Other

Data Science, EDA, Geospatial Analytics, GeoPandas, Data Analysis, Visualization, Metabase, Spatial Analysis, Data Science, Data Visualization, Dash, Statistics, Machine Learning, FastAPI, QGIS, Time Series, System Design, Principal Component Analysis (PCA), Web Scraping, Geospatial Data, K-means Clustering, Simulations, NLP, Support Vector Machines (SVM), SVMs, Benchmarking, Generative Pre-trained Transformers (GPT), Deep Learning

Collaboration That Works

How to Work with Toptal

Toptal matches you directly with global industry experts from our network in hours—not weeks or months.

1

Share your needs

Discuss your requirements and refine your scope in a call with a Toptal domain expert.
2

Choose your talent

Get a short list of expertly matched talent within 24 hours to review, interview, and choose from.
3

Start your risk-free talent trial

Work with your chosen talent on a trial basis for up to two weeks. Pay only if you decide to hire them.

Top talent is in high demand.

Start hiring