Methodology & caveats

CycloTracker maps reported Cyclospora cayetanensis cases by U.S. state, focused on the most recent ~3 months and refreshed weekly. This page explains exactly what the numbers are — and, just as importantly, what they are not.

What is counted

Case counts come from the CDC NNDSS weekly tables (nationally notifiable disease surveillance). These are provisional reports by week and jurisdiction, revised in later weeks. They are surveillance case reports — not lab-confirmed-only figures and not outbreak-linked counts (those are a different unit; see NORS below).

The 3-month window

The default view sums each state's weekly counts over the latest 13 MMWR weeks present in the data. The year selector instead shows a year's cumulative year-to-date total. A state can show data in one view and not the other.

Zero vs. no data — never conflated

A blank is never treated as a zero. Each state-week carries a status:

• reporting — a real count (may be zero, shown as an empty state, no symbol);
• no data — the disease is flagged not-notifiable (N) or unavailable (U) for that jurisdiction. These states are drawn with a hatched fill, never a zero-sized symbol. In the current window this affects Idaho, Mississippi, and Pennsylvania.

Comparability caveats

Cyclosporiasis case definitions change over time (see CSTE position statements), and reporting completeness varies by state and season. Year-over-year and state-to-state comparisons should be read with that in mind. Provisional weekly counts are revised. Domestically acquired summer outbreaks (often produce-linked) drive large, real spikes — the current Ohio/Michigan surge is one.

Reconciliation & precedence

Where multiple sources overlap, the value is chosen by a documented precedence order (state DOH > NNDSS annual > NNDSS weekly); only sources with a usable number can win, and disagreements are flagged rather than averaged. Today NNDSS weekly is the single enabled case source, so nothing overlaps yet — the layer exists so additional sources slot in cleanly.

Rates (per 100,000)

Per-capita rates use U.S. Census ACS 1-year population (variable B01001_001E). The brief called for the Census PEP API, but PEP's population endpoint stops at Vintage 2021; ACS 1-year is current (2024). A rate is only shown where population is known — never fabricated.

Symbol encoding

Circle area is proportional to the value (radius scales with the square root) — so a state with 4× the cases gets a circle 2× the radius, not 4×. New York City is reported separately by NNDSS and is summed into New York State. Territories are stored but not drawn (Albers USA shows the 50 states + DC).

What we never do

• Never fabricate or interpolate case data — missing is missing.
• Every rendered number is traceable to a source (see the table below and each tooltip).
• Never scale symbols by radius; never render "no data" as zero.

Sources