How utsapoddar/msds-time-estimator works
MSDS Time Estimator turns review observations and a learner’s completed-course history into personalized course-duration and degree-total estimates, with uncertainty intervals. The main workflow refreshes review data, builds Bayesian course-hour posteriors, uses anchor courses and learner inputs to estimate pace, then predicts target-course durations and aggregates a selected degree plan. A separate simplified predictor offers a lighter history-ratio alternative; the course decision matrix supports ranking courses. The graph follows documented capabilities and sampled code; some links to unsampled implementations are omitted where call ownership is unclear.
Main entry points: pipeline.py Prediction pipeline · form_intake.py Review intake · simplified_predictor.py Simplified predictor · course_params.csv Course parameters
Components 18
Review intake
- Review intake
modules/form_intake.py - Review observations
data_points.csv - Pending reviews
data_points_review.csv
Estimation model
- Course parameters
course_params.csv - Bayesian updates
modules/bayesian_priors.py - Prediction pipeline
modules/pipeline.py - Skill factors
modules/latent_factors.py - Topic similarity
modules/course_topics.py - Focus adjustment
modules/focus_ratio.py
Predictions
- Time distributions
modules/lognormal_marginals.py - Course estimates
modules/pipeline.py - Degree total
modules/pipeline.py - Example run
example_run.py
Alternatives
- Simplified predictor
modules/simplified_predictor.py - Guide priors
modules/guide_priors.py - Course ranking
course_decision_matrix.py
Other components
- Learner
- Published reviews
Connections 19
- Published reviews → Review intake: provides reviews
- Review intake → Review observations: refreshes data
- Review intake → Pending reviews: queues submissions
- Learner → Prediction pipeline: supplies profile
- Course parameters → Prediction pipeline: loads parameters
- Review observations → Prediction pipeline: loads observations
- Prediction pipeline → Bayesian updates: updates posteriors
- Prediction pipeline → Skill factors: estimates skills
- Prediction pipeline → Topic similarity: transfers pace
- Prediction pipeline → Focus adjustment: adjusts schedule
- Prediction pipeline → Time distributions: builds intervals
- Prediction pipeline → Course estimates: produces estimates
- Prediction pipeline → Degree total: aggregates plan
- Example run → Prediction pipeline: runs predictions
- Learner → Simplified predictor: supplies history
- Course parameters → Simplified predictor: provides baselines
- Guide priors → Simplified predictor: provides difficulty fit
- Learner → Course ranking: tunes priorities
- Course estimates → Course ranking: informs workload