Griffith Consulting Group
Book a call

data engineering · stage 02 of 04

Data cleaning, so the counts come out right

One customer shows up twice. One channel carries three spellings. Prices arrive as text. We fix those slips before anyone counts anything, and we show you each change.

Book a call See an example
namechannelpriceAna RuizMarcus BellReferral95referralReferral$6464Marcus Bell␣95REFERRALReferralreferal48.00Gabe MossReferral48count 4 · sum errorcount 3 · sum 207one spelling per channelprices as numbersone row per customer
fig. 1 · one sheet, cleaned: one spelling per channel, prices as numbers, one row per customer

What you get

  • One label per channel

    Walk in, walk-in and Walk-In become one label, so each channel gets one count.

  • Duplicates merged

    One customer, one row, even when a name came in with a stray space.

  • Numbers you can add

    $64, 95 and 48.00 become 64, 95 and 48, so totals and averages work.

How it works

  1. Look before fixing

    We count the blanks, spellings and repeats in each column.

  2. Agree on the rules

    You tell us which spelling wins and what counts as the same customer.

  3. Apply the fixes

    A script applies the same rules to every row.

  4. Check the totals

    We compare the counts before and after, and show you what changed.

A worked example

Sample records from a fictional bakery, the same one that runs through the overview.

  1. Data cleaning: fix the mess

    One customer can show up twice, and one channel can carry three spellings. Left alone, those slips give wrong counts. We standardize labels, merge duplicates and turn prices into numbers.

    you get counts you can trust

    • walk inWalk-inone label
    • Marcus Bell with a trailing spaceMarcus Bellduplicate merged
    • $64 · 95 · 48.0064 · 95 · 48prices as numbers

Questions owners ask

Have another one? Ask it on a call.

What do you fix?

Duplicate customers, labels with several spellings and prices stored as text.

How do I know what you changed?

We show you each fix as a before and after, like the list on this page.

What happens after cleaning?

Analysis. We answer your business question with the cleaned numbers.

The other stages

Each stage feeds the next: mining, cleaning, analysis, then visualization.

Tell us what's piling up

Describe the mess in a few sentences. We'll reply with where we'd start.

Book a call