Data work swings between two very different moods. One is careful reasoning about how tables relate, where a join can silently duplicate rows, and what a late-arriving file will do to yesterday’s totals. The other is plain upkeep: rotating credentials, clearing temp storage, renaming columns. If you treat both the same, you will do one badly. This checklist helps you match tasks to energy levels, so reasoning-heavy pipeline work gets your clear hours and the upkeep gets the rest.
Treat the list as a Friday-afternoon audit of how your week actually ran. Go through each item, answer yes or no, and close the gaps in order. Nothing here needs new tools.
Part one: know your curve
- Have you recorded your alertness for at least five working days? A quick one-to-five rating at set times is enough. Without it, you are guessing.
- Can you name your peak window and your slump window? Write them down as clock ranges. Vague answers like “mornings” are too loose to schedule against.
- Do you know what shifts your curve? Short sleep, heavy meals and a run of meetings all flatten people. Note which ones affect you so you can plan around them.
Part two: sort the work you actually do
Make a list of everything that regularly lands on your plate. For a data engineer, it usually looks something like this. Mark each as sharp, steady or slow.
- Designing a new model layer or partitioning scheme: sharp.
- Working out why a metric disagrees between two reports: sharp.
- Reviewing a colleague’s transformation logic: steady.
- Responding to routine access requests: slow.
- Running a scheduled backfill and watching it finish: slow.
- Tidying orphaned tables and old test datasets: slow.
Audit question: is anything in the sharp list currently living in your slump hours? Most people find at least two items in the wrong place. Those are your first moves.
A worked example helps. Suppose a reconciliation job has started producing totals that are slightly off every Monday. Finding the cause means holding several ideas in mind at once: time zones, late data, a changed upstream field. That is sharp work and belongs in the peak. Once the cause is clear, the fix may be a one-line change plus a rerun, which is slow work and can wait for the afternoon. Splitting the job like that is the whole habit in miniature.
Part three: the pipeline-design check
- Is design work booked into a named peak slot? If it only happens “when there is time”, it will be squeezed into whatever hour is left, which is usually the worst one.
- Is your chat status set while you work on it? A hard design problem does not survive three interruptions. Treat the peak as a scarce resource.
- Do you finish with a written next step? Ending a design session with a note about the open question saves the first twenty minutes of tomorrow.
Part four: the housekeeping check
- Is there a visible list of slump-friendly jobs? Clearing old partitions, refreshing documentation links, closing stale tickets. When your focus drops, you pull from the list without having to decide.
- Are those jobs low-risk? Do not schedule destructive operations, such as dropping tables or deleting storage, for the tired hour. Move anything irreversible to a time when you can read each line properly. If a chore needs a careful review before running, it is not a slump task.
- Do you have a stopping rule? Housekeeping can swallow a whole afternoon. Set a limit and leave the rest for another day.
Part five: the review and handover check
- Are code reviews placed in the middle of your curve? Reviews need attention but not originality, so they fit the shoulder hours well. A tired review misses things, and a peak-hour review wastes your best time.
- Do you write handover notes while you still have some energy? Many people leave this to the very end and produce a useless one-liner. A five-minute note written at the tail of your afternoon is far more useful to the next person.
For an example of the sort of work that benefits from a clear head, see this walk-through on keeping live analytics flowing when pipelines fail. Reading about breakage is one thing, but diagnosing it yourself at the wrong hour is another.
One more thing to check here is how you treat the in-between jobs. Things like writing a data-quality test or updating a data dictionary are neither hard nor trivial. Place them in the shoulders of your day, after the peak starts to fade but before the slump bites. If you notice you are rereading the same sentence of a test three times, that is your signal to switch to the slow list rather than push on.
Part six: the team-level check
- Does your team know each other’s peak windows? Even a rough idea helps. You can then avoid dropping a complex request on someone at the end of their day.
- Are on-call and release duties spread so no one is always drained? Energy matching is harder when the same person always carries the late shift.
- Are routine requests batched into a predictable slot? If a colleague knows that access requests are handled mid-afternoon, they stop pinging you in the morning.
Part seven: the civic-minded side note
Outside work, the same audit applies to small personal admin jobs. If you are drafting a letter about a local concern, do it in a sharp hour. If you are curious where people post such letters, JanBoard is one independently run example. It describes itself as sitting outside government and leaves formal grievance and legal channels untouched. Whether you use it or not, the scheduling principle holds.
Scoring yourself without overthinking it
Count your yes answers. If you got most of them, your week already respects your energy and only needs small tweaks. If you got about half, pick the two or three gaps that annoy you most and fix those first. If you got very few, do not try to overhaul everything. Start with part two: move one sharp task into a peak slot and one slow task into a dip slot, and see how the next three days feel.
Keep it boring and repeatable. The point of a checklist is that you do not need to remember it. Put the short version on a sticky note or in the top of your task list: peak equals design and diagnosis, middle equals review and talk, slump equals upkeep.
Re-run the audit every month
Energy patterns drift with the seasons, with changes in your home life and with the shape of your job. A quick monthly pass through the checklist keeps it honest. If something no longer fits, such as a new meeting that sits across your peak, treat it as a scheduling problem to solve rather than a personal failing.
One more tip is to read a small related guide when you are in the slow lane. The article about running data operations across several clouds is the sort of reading that suits a steady-energy hour, because it asks for comprehension rather than invention. Save the creative, high-stakes design for when you are sharp, and let the lighter material fill the gaps.