Tableau or TensorFlow First: What Beginners Should Pick

// Tableau calculated field: "Margin Status"
// (Analysis > Create Calculated Field)
IF SUM([Profit]) / SUM([Sales]) < 0.10 THEN "Low margin"
ELSEIF SUM([Profit]) / SUM([Sales]) < 0.25 THEN "Healthy"
ELSE "Strong"
END

// Use it:
// 1. Drag [Region] to Rows
// 2. Drag SUM([Sales]) to Columns
// 3. Drag [Margin Status] to Color
// Result: a bar chart showing which regions have weak margins

# pip install tensorflow scikit-learn
import tensorflow as tf
from sklearn.datasets import load_iris
from sklearn.model_selection import train_test_split

X, y = load_iris(return_X_y=True)
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=1)

model = tf.keras.Sequential([
    tf.keras.layers.Input(shape=(4,)),
    tf.keras.layers.Dense(16, activation="relu"),
    tf.keras.layers.Dense(3, activation="softmax"),
])
model.compile(optimizer="adam",
              loss="sparse_categorical_crossentropy",
              metrics=["accuracy"])
model.fit(X_train, y_train, epochs=60, verbose=0)

loss, acc = model.evaluate(X_test, y_test, verbose=0)
print("test accuracy:", round(acc, 3))

Analyst roadmap (Tableau first)

Stage 1  Excel: formulas, pivot tables, charts
Stage 2  SQL: SELECT, JOIN, GROUP BY
Stage 3  Tableau: connect data, basic charts
Stage 4  Tableau: calculated fields, filters, dashboards
Stage 5  Storytelling: one dashboard, one clear message
Stage 6  Portfolio: 3 dashboards on public data
Stage 7  Python basics (optional next step)

AI roadmap (TensorFlow later)

Stage 1  Python fundamentals
Stage 2  NumPy and pandas
Stage 3  Statistics and maths basics
Stage 4  Classic ML with scikit-learn
Stage 5  Neural network concepts
Stage 6  TensorFlow / Keras: build and train models
Stage 7  Projects: images or text, with evaluation

Good beginner dashboard checklist

[ ] One clear question, e.g. "Which regions are falling behind?"
[ ] 3-5 charts, not 15
[ ] Key numbers (KPIs) at the top
[ ] Filters for date and region
[ ] Consistent colours, red only for problems
[ ] Titles that state the insight, not just the metric
[ ] Tested on a smaller screen
[ ] One-paragraph summary of findings

Model Training Log (for later, with TensorFlow)

Run  Layers     Epochs  Train acc  Test acc  Notes
1    16         60      ____       ____      baseline
2    32-16      60      ____       ____      deeper
3    32-16      120     ____       ____      longer

Compared with a simpler scikit-learn model?  Y / N
Overfitting signs? _______________

Tableau or TensorFlow - Decision Sheet

My goal role: analyst / BI / data scientist / ML engineer / unsure
Do I know Excel well?    Y / N
Do I know SQL?           Y / N
Do I know Python?        Y / N
Do I know ML basics?     Y / N

Mostly "analyst/BI" or "unsure"  -> Tableau first
"ML engineer" + Python + ML      -> TensorFlow next
Otherwise -> build Python and ML basics first

Decision: _______________

Dashboard Brief

Audience: _______________
Main question: _______________
Data source: _______________
KPIs: 1. ___  2. ___  3. ___
Charts planned: _______________
Filters: _______________
Insight summary (after building): _______________

Monthly Progress Log

Tool focus: Tableau / Python / TensorFlow
Hours practised: ____
Built: _______________
Shared publicly?  Y / N
Feedback received: _______________
Next month's goal: _______________