projects.html

<!DOCTYPE html>
<html>
<head>
    <title>Converted from: projects.md</title>
    <meta charset="utf-8"/>

</head>
<body>

<div
            class="github-markdown-body"
            data-color-mode="auto"
            data-light-theme="light"
            data-dark-theme="dark"
        >
            <div class="github-markdown-content"><span id="markdown-mermaid" aria-hidden="true"
                    data-dark-mode-theme="dark"
                    data-light-mode-theme="default"></span>
                <table data-line="0" class="code-line" dir="auto">
<thead data-line="0" class="code-line" dir="auto">
<tr data-line="0" class="code-line" dir="auto">
<th data-line="0" class="code-line" dir="auto">layout</th>
<th data-line="0" class="code-line" dir="auto">title</th>
<th data-line="0" class="code-line" dir="auto">subtitle</th>
<th data-line="0" class="code-line" dir="auto">comments</th>
<th data-line="0" class="code-line" dir="auto">carbonads</th>
</tr>
</thead>
<tbody data-line="0" class="code-line" dir="auto">
<tr data-line="0" class="code-line" dir="auto">
<td data-line="0" class="code-line" dir="auto">page</td>
<td data-line="0" class="code-line" dir="auto">Streamlining Data Science</td>
<td data-line="0" class="code-line" dir="auto">Exploring the Frontiers of Data Analysis</td>
<td data-line="0" class="code-line" dir="auto">true</td>
<td data-line="0" class="code-line" dir="auto">true</td>
</tr>
</tbody>
</table>
<div  data-line="8" class="code-line" dir="auto" ></div>
<style>
.project-section { margin-bottom: 30px; }
.project-title { color: #2c3e50; border-bottom: 2px solid #3498db; padding-bottom: 5px; }
.project-description { margin-left: 20px; }
.project-list { list-style-type: none; padding-left: 0; }
.project-list li { margin-bottom: 10px; }
.project-list li:before { content: "▹"; color: #3498db; margin-right: 10px; }
.codebase-link { background-color: #3498db; color: white; padding: 5px 10px; border-radius: 5px; text-decoration: none; }
</style>
<h1 id="heading-data-science-portfolio">Data Science Portfolio</h1>
<p>As a data scientist, I believe in the power of open source, open science, and open data. My recent focus has been on productionizing machine learning and deep learning tools for versatile use across various datasets.</p><h2 id="heading-current-projects">Current Projects</h2>
<h3 id="heading-a.-data-science-for-bioinformatics-28omixhub29">A. Data Science for Bioinformatics (OmixHub)</h3>
<p><a href="markdown-viewer.html?file=test.md" target="_blank">Test Markdown File</a></p><div  data-line="26" class="code-line" dir="auto" ></div>
<div class="project-section">
<h3 class="project-title">Vision</h3>
<ul class="project-list project-description">
  <li>Build well-documented, user-friendly modules for omics datasets</li>
  <li>Provide tutorials for early-stage data scientists in bioinformatics</li>
  <li>Standardize ML/DL and bioinformatics tools</li>
</ul>
<div  data-line="34" class="code-line" dir="auto" ></div>
<h3 class="project-title">Implemented Projects</h3>
<ul class="project-list project-description">
  <li><a href="markdown-viewer.html?file=https://abhilashdhal.com/files/DeSeqApplication/docs/workflow.md">Differential Gene Expression in Kidney Cancer Subtypes</a></li>
  <li><a href="markdown-viewer.html?file=https://abhilashdhal.com/files/SupervisedLearningApplication/docs/workflow.md">Multi-label Classification of Kidney Cancer Subtypes</a></li>
  <li><a href="markdown-viewer.html?file=https://abhilashdhal.com/files/SuperviseLearningEnsembleApplication/workflow.md">Ensemble Classifiers for Kidney Cancer Classification</a></li>
  <li><a href="markdown-viewer.html?file=https://abhilashdhal.com/files/UmapApplication/docs/workflow.md">Disease Stratification using UMAP and DBSCAN</a></li>
</ul>
<div  data-line="42" class="code-line" dir="auto" ></div>
<p><strong>Data source:</strong> <a href="https://portal.gdc.cancer.gov/repository">Genomic Data Commons</a></p>
<p><strong>Codebase:</strong> <a href="https://github.com/adhal007/OmixHub" class="codebase-link">OmixHub</a></p>
</div>
<h3 id="heading-b.-data-science-for-e-commerce">B. Data Science for E-Commerce</h3>
<div  data-line="48" class="code-line" dir="auto" ></div>
<div class="project-section">
<h3 class="project-title">Vision</h3>
<ul class="project-list project-description">
  <li>Develop generalizable tools for e-commerce analysis</li>
  <li>Demonstrate advanced Python programming (OOP) with real-world data</li>
</ul>
<div  data-line="55" class="code-line" dir="auto" ></div>
<h3 class="project-title">Projects</h3>
<ul class="project-list project-description">
  <li><a href="/files/E_commerce_projects/customer_segmentation_model.md">Customer Segmentation using K-means Clustering</a></li>
  <li><a href="/files/E_commerce_projects/groupon_deals_data_analysis.md">Analysis of Groupon Deals using Propensity Scores</a></li>
</ul>
<div  data-line="61" class="code-line" dir="auto" ></div>
<p><strong>Data source:</strong> Kaggle</p>
<p><strong>Codebase:</strong> <a href="https://github.com/adhal007/Data-Science-with-Python" class="codebase-link">E-Commerce Projects</a></p>
</div>
<h3 id="heading-c.-data-science-for-medical-diagnosis">C. Data Science for Medical Diagnosis</h3>
<div  data-line="67" class="code-line" dir="auto" ></div>
<div class="project-section">
<h3 class="project-title">Vision</h3>
<ul class="project-list project-description">
  <li>Assemble deep learning algorithms for medical imaging analysis</li>
  <li>Create unified modules for 2D and 3D image processing</li>
</ul>
<div  data-line="74" class="code-line" dir="auto" ></div>
<h3 class="project-title">Projects</h3>
<ul class="project-list project-description">
  <li><a href="/files/ComputerVision/imbalanced_xray_classification.md">Multi-label Classification of X-Rays using DenseNet</a></li>
</ul>
<div  data-line="79" class="code-line" dir="auto" ></div>
<p><strong>Data source:</strong> Contact via <a href="mailto:adhalbiophysics@gmail.com">email</a></p>
<p><strong>Codebase:</strong> <a href="https://github.com/adhal007/AIMedicine" class="codebase-link">AI for Medicine</a></p>
</div>
<h3 id="heading-d.-grad-school-projects">D. Grad School Projects</h3>
<div  data-line="85" class="code-line" dir="auto" ></div>
<div class="project-section">
<p>During my masters at <a href="https://cs.ucdavis.edu/">UC Davis</a>, I developed skills in quantitative genetics and AI:</p>
<div  data-line="88" class="code-line" dir="auto" ></div>
<ul class="project-list">
  <li><a href="https://github.com/adhal007/SSBR-JWAS-Implementation">Bayesian Regression Models in Animal Genomics</a></li>
  <li><a href="https://www.overleaf.com/1359439834vszcmwdsjymd#b028db">Deep Learning for Intracranial Hemorrhage Detection</a></li>
  <li><a href="https://github.com/adhal007/SC2-GCP-CNN">Real-time Strategy Agents for Starcraft II</a></li>
</ul>
</div>
<div  data-line="95" class="code-line" dir="auto" ></div>
<!-- ### Software Development /Object Oriented Programming (OOP) Practices:
I'm a big advocate of writing re-usable and streamline code for data science projects. Some of the modules that I've built for my personal bioinformatics projects can be found on [OmixHub](https://github.com/adhal007/OmixHub). Some of the crucial and advanced concepts of **inheritance, polymorphism, encapsulation and abstraction** can be seen in all of the modules.

Here is a list of some useful tools that can be found:

- **[Base Preprocessor](https://github.com/adhal007/OmixHub/blob/main/src/base_preprocessor.py)**
  - Example class with methods and attributes inherited by child classes. Some key methods provided are:
    - data skew
    - data leakage
    - patient overlap
    - training_testing_split 
    - etc
- **[Base ml models wrapper](https://github.com/adhal007/OmixHub/blob/main/src/base_ml_models.py)**
  - Example class with functionality to evaluate and plot multiple ML models for a data science application. This is intended to be inherited by specific child classes for building models for different Omics data
- **[Dimensionality reduction and clustering wrapper](https://github.com/adhal007/OmixHub/blob/main/src/DimRedMappers/README.md)**
  - Interfaces with Preprocessor classes to facilitate easy application of UMAP and clustering
- **[Differential Analysis Wrapper](https://github.com/adhal007/OmixHub/blob/main/src/pydeseq_utils.py)**
  - Faciliatates easy application of pydyseq in a few lines to perform differential analysis -->
</div>
        </div>

</body>
</html>