Looking back for all the studies in past year, here comes a high level understanding/summary to the Data Science process - where how far we’ve gone and we’ll start in next year.

Stages

Stage 1: Business Goal Understanding

1.1 what is the problem?
1.2 what is the goal?
1.3 how can you use data analytics to achieve the goal?

Stage 2: Data Acquire

2.1 what kinds of data are needed?
2.2 what data you have and what you don’t?
2.3 how to identify data sources, collect and integrate data?

Stage 3: Data Preparation

3.1 exploratory

• stats (mean, median, max, min, skewness, kurtosis etc.)
• Viz (hist, scatter, box, bar, line, heatmap etc.)

3.2 pre-process

• data clean (missing, duplicate, invalid, outlier etc.)
• feature engineering (add, remove, combining, breaking etc.)
• transformation (scaling, normalization, dimensionality reduction etc.)

Stage 4: Data Analysis

4.1 select analytic approaches
4.2 build models
4.3 assess results

Stage 5: communicate results, feedback, improvements

Tableau public, Rpubs etc. tools may be useful,

Stage 6: apply results, feedback, improvements

And of cause, the whole process consists of iterative cycles of each stage.

Tools that are usually involved:

Tools: Hadoop Ecosystem, PostgreSQL, MongoDB, Cassandra, Neo4j, Tableau, Splunk, Elastic Stack, KNIME etc.
Platforms are useful: Kaggle, BDU Labs, databricks, Cloudera Quickstart etc.

Ref: Coursera, BDU etc.