←back to projects
DBS Foundation Coding Camp 2024/AI/ML/course

Books Recommendation System

Built as a machine-learning coursework artefact, this project documents the business framing, dataset choice, notebook implementation, and evaluation discussion behind a book recommendation workflow.

project links
Repositoryrepo
Domain
AI/ML
Role
Machine Learning Engineer
Output
ML Pipeline
Category
Recommendation System
Project Framing

A source-backed case study built for recruiter review

This reading path makes the problem choice, evidence quality, user framing, execution decisions, and proof trail visible without overstating what the sources support.

Project Type
course

Recommendation-system project comparing content-based and collaborative filtering approaches for book discovery.

Orientation
Tech

Demonstrates recommender-system reasoning through business understanding, data preparation, modeling alternatives, and evaluation narrative.

Core Stack
Python · Jupyter Notebook · TensorFlow · Pandas

Notebook-based ML workflow combining book metadata and rating data with content-based similarity logic and collaborative filtering model experimentation.

Why This Problem Mattered

Problem framing before execution

The case-study layer starts with why this problem was selected and how the context justified investment.

Problem Framing Map

Issue

Readers struggle to discover relevant books when catalog metadata and user-rating signals are not translated into usable recommendation logic.

Context

The project was built as a coursework recommendation artefact that combines business framing, dataset choice, and comparative modeling approaches in a notebook workflow.

Why Selected

It strengthens the portfolio by showing that recommendation projects require product reasoning, data preparation, and model tradeoff explanation together.

Problem statement

Readers can struggle to discover relevant books when catalog and rating data are not transformed into personalized recommendations.

Solution thesis

Implemented and documented recommendation approaches using content-based filtering and collaborative filtering on the Kaggle Book Recommendation Dataset.

Research and Evidence

What supports the narrative

Evidence is surfaced with its source type and credibility note so the recruiter can quickly see what is directly backed versus intentionally constrained.

Approach comparison
local

The README documents both content-based filtering and collaborative filtering approaches.

Credibility: Directly supported by the project metrics and README-backed sources.
Dataset framing
local

The source explicitly identifies the Kaggle Book Recommendation Dataset as the data foundation.

Credibility: Backed by the project metrics and notebook-aligned project record.
Review artefacts
local

The local archive includes README, notebook, notebook export, and evaluation image artefacts.

Credibility: Supported by the current metrics and proof sections of the project entry.

Credibility Notes

  • ●The project is presented as recommendation-system reasoning and experimentation, not as a deployed personalized reading platform.
  • ●No real user click-through, retention, or conversion impact is claimed beyond the notebook-backed evaluation story.
Who The User Was

User framing stays explicit

When formal research artefacts are not available, the page still explains who the work served and why that user framing is justified by the existing sources.

Primary user
Readers who need better discovery support across large book catalogs.

The project problem statement and solution framing explicitly focus on recommendation usefulness for discovery.

Reviewer stakeholder
Technical reviewers comparing how metadata-driven and preference-driven recommendation logic are reasoned about.

A major source-backed strength of the project is the side-by-side explanation of multiple recommender approaches.

Decision Flow

How design thinking translated into decisions

The goal is to show the trace from research and insight to concrete product or system decisions, then to the outcomes those decisions supported.

Design Thinking Flow

Each step keeps the movement from evidence to action explicit before the rationale expands it.

  1. Step 1
    Discovery problem framing

    Started from the user problem of book discovery rather than from algorithm choice alone.

    Signal: Recommendation quality is tied to product need, not only model mechanics.
  2. Step 2
    Approach comparison

    Used both content-based and collaborative filtering to expose different recommendation tradeoffs.

    Signal: The project becomes a reasoning exercise, not just a single-model implementation.
  3. Step 3
    Traceable notebook narrative

    Kept the README and notebook aligned so business framing, data handling, and evaluation discussion remain reviewable.

    Signal: The artefact trail supports recruiter-safe technical discussion.

Decision Rationale

Each decision keeps the path from insight to execution visible before ending on the outcome signal.

Dual recommender framing
Insight

Recommendation quality depends on whether the system should reason from item similarity, user preference patterns, or both.

Decision

Included both content-based filtering and collaborative filtering approaches.

Outcome

The project surfaces recommender tradeoffs more clearly than a single-method notebook would.

Notebook-first delivery
Insight

The main project value lies in making business framing, data preparation, and model comparison easy to inspect together.

Decision

Used a notebook-centred workflow with README alignment and evaluation artefacts.

Outcome

The recommender story stays reviewer-traceable without overstating deployment maturity.

Solution and System Execution

Execution choices and delivery details

This section preserves the technical and operational substance: architecture, responsibilities, trade-offs, and implementation quality signals.

System Design

Notebook-based ML workflow combining book metadata and rating data with content-based similarity logic and collaborative filtering model experimentation.

Source-backed Impact

Demonstrates recommender-system reasoning through business understanding, data preparation, modeling alternatives, and evaluation narrative.

Responsibilities

  • ●Framed recommendation problem statements and goals
  • ●Prepared book and rating data for recommendation experiments
  • ●Documented evaluation considerations for both recommendation approaches

Stack Decisions

  • ●Used content-based filtering to reason from item attributes
  • ●Used collaborative filtering to incorporate user-rating preference signals
  • ●Kept notebook and README narrative aligned for reviewer traceability

Trade-offs

  • ●Used a notebook-centred delivery format rather than a deployed recommendation service
  • ●Presented evaluation as source-backed project evidence without adding unsupported user-impact claims

Challenges

  • ●Balancing item-metadata signals with rating-based preference signals
  • ●Explaining recommender tradeoffs clearly enough for portfolio review
Outcomes and Proof

What was delivered and what can be verified

Outcome claims remain conservative and source-backed, while proof records and recruiter-safe links surface the strongest verification trail available.

Validation Signals

  • ●README documents both content-based and collaborative filtering approaches.
  • ●Local archive includes README, notebook, notebook export, and evaluation image artefacts.

Source-backed Outcomes

  • ●README documents content-based filtering and collaborative filtering approaches
  • ●Local archive includes README, notebook, notebook export, and evaluation image artefact
  • ●Dataset source documented as Kaggle Book Recommendation Dataset
Retrospective and Limits

What the project proves, and what it does not

Strong case studies show both what was learned and where the current evidence stops.

Retrospective

Next iteration should add environment setup instructions and a compact reproducibility checklist for the notebook run.

Evidence Limits

  • ●Current sources do not support live product metrics, online recommendation feedback, or deployed personalization outcomes.
  • ●The project should remain framed as recommendation-system experimentation and evaluation reasoning.

Lessons

  • ●Recommendation quality should be tied to the product discovery problem, not only model mechanics
  • ●Combining content and collaborative perspectives makes recommender tradeoffs easier to evaluate