Day1_​Health_​Check

1 · Data Access

Read the fleet feed and alarm log from file.

2 · Data Cleaning

Make both feeds trustworthy.

3 · Data Transformation

Derive the deviation signal and flag it.

4 · Data Merging

Join alarms onto the readings; deduplicate.

5 · Data Aggregation

Summarise into a per-pump health table.

6 · Export & Visualisation

Chart the fleet, then write the Day-1 report.

Step 1a · Read the data

Read the fleet sensor feed (17,472 rows) with the CSV Reader node

Step 2a · Clean the fleet feed

1. Convert the timestamp to a date with the String to Date&Time node
2. Keep only the trusted time window with the Row Filter node
3. Standardise the pump_id text with the String Manipulation node
4. Sort by pump, then time with the Sorter node
5. Fill the 140 missing temperature gaps with the Missing Value node

Step 3 · Transform

1. Compute deviation = temp - 60 with the Math Formula node
2. Flag the status (666 = HIGH) with the Rule Engine node

Step 4 · Merge

1. Left-join the alarms onto the readings with the Joiner node
2. Remove duplicate rows with the Duplicate Row Filter node

Step 5 · Aggregate

1. Build the per-pump health table with the GroupBy node
2. Compute deviation vs fleet mean with the Math Formula node
3. Mark REVIEW when |dev| > 2 with the Rule Engine node

Step 6a · Visualise

1. Give each pump one colour with the Color Manager node
2. Show behaviour over time with the Line Plot node
3. Show the fleet distribution with the Histogram node
4. Show spread per machine with the Box Plot node

Step 1b · Read the alarm log

Read the alarm log (192 rows) with the CSV Reader node

Step 2b · Clean the alarm log

1. Parse the alarm time with the String to Date&Time node
2. Clean the alarm pump_id with the String Manipulation node

Step 6b · Bar chart + report

1. Show the fleet at a glance with the Bar Chart node
2. Write the Day-1 report with the CSV Writer node