Jamel Dargan

• Presented results of a display advertising campaign, calculated ROI, compared Ad Group performance, and performed keyword analysis.
• Compared metrics, including click-through-rate, conversion rate, and cost-per-acquisition and their weighted value, based on campaign objecives.
• Highlighted the potential to effect performance via A/B testing and keyword matching.

• Audited an existing, provided site for optimization.
• Used industry tools from Moz, SEMRush, Pingler, and Google, to identify a number of opportunities to optimize metadata, pursue quality backlinks, and to improve site speed.

• Created a target objective for to generate 3 leads in 3 days, generating positive ROI, on a $125 budget.
• Achieved positive ROI but missed its lead-generation target, highlighting the challenges of short-term campaigns and the usefulness of ad variant tests.
• Suggested ways forward, regarding demographics and location (geo-fencing), based on Ad Set performance results.

Debugging a deep learning model, using the Pytorch Influence Functions package | Which training data exert the most influence over a neural network model's outcomes?

framework: Jupyter Notebook
languages: Python

libraries and modules:
CUDA
Numpy
OS
PyTorch:
pytorch-influence-functions
torchvision
torch.nn
torch.optim
Sys

visualization libraries:
Matplotlib

Binary classification using sklearn pipeline, sequential feature selection, multiple algorithm comparison, hyperperameter tuning, and XGBoost | What features are best able to predict how often an online article will be shared?

framework: Jupyter Notebook
languages: Python

libraries and modules:
Mlxtend SequentialFeatureSelector,
Numpy,
OS,
Pandas,
Pickle,

Scikit-Learn:
ensemble,
linear_model,
metrics,
model_selection,
pipeline… Show more

Binary classification using sklearn pipeline, sequential feature selection, multiple algorithm comparison, hyperperameter tuning, and XGBoost | What features are best able to predict how often an online article will be shared?

framework: Jupyter Notebook
languages: Python

libraries and modules:
Mlxtend SequentialFeatureSelector,
Numpy,
OS,
Pandas,
Pickle,

Scikit-Learn:
ensemble,
linear_model,
metrics,
model_selection,
pipeline,
preprocessing,
tree,
XGBoost,
Yellowbrick classifier,
DiscriminationThreshold,
roc_auc,

visualization libraries:
Matplotlib,
Seaborn,
Mlxtend plotting Show less

Exploratory data analysis of recently successful movie titles, genres, and studios.

framework: Jupyter Notebook
languages: Python

Libraries and modules, incl.:
Pandas,
Numpy,
Sqlite3,

Visualization:
Matplotlib,
Seaborn