Design note. Portable: nothing here is specific to one sensor deployment. Target reader is a scientifically literate engineer who is not a Kalman-filter specialist.
Sensor series arrive irregularly: a nominal cadence with jitter, dropped packets, hour-long radio outages, occasional multi-day gaps. Downstream consumers (charts, tables, alerting, correlation against other series) want a regular grid — say 10-minute steps — and want to know how much to
| Welcome to the battery doctor's cave sermon. Man, we've got such juicy | |
| material to talk about today, no joke. Our donut lab came out and | |
| published an all solid state battery, a completely solid cell. And | |
| buddy, let me tell you, I've been asked so many times whether this is | |
| real, whether this can actually happen, whether something like this | |
| can come out of Finland, some donut company that's gonna free us all | |
| from misery and be the new Nokia. What would this actually mean? This | |
| solid battery has definitely generated a lot of skepticism in the | |
| media, lots of it. Lots of hope, like yeah, this is definitely going | |
| to happen, this is going to be a huge thing for Finland in the future, |
| data { | |
| int<lower=0> N; // rows | |
| int<lower=1> D; // delays, first is 0, last D-1 | |
| int<lower=0> T; // days, first is 1 | |
| int<lower=0> counts[N]; | |
| int<lower=1> t[N]; | |
| int<lower=0, upper=D-1> dt0[N]; | |
| int<lower=0, upper=D-1> dt1[N]; | |
| real<lower=0> dir_prior; |
| data { | |
| int<lower=1> n; // n = 2632 | |
| int y[n]; // number of shootings per day | |
| } | |
| transformed data { | |
| int<lower=1,upper=365> yday[n]; | |
| for (i in 1:n) yday[i] = i % 365 + 1; | |
| } | |
| parameters { | |
| vector[n] mu_innovations; |
| library(dplyr) | |
| # install.packages("rgeos", type="source") | |
| # install.packages("rgbif", type="source") | |
| # library(rgbif) | |
| # d <- occ_search(orderKey=797, datasetKey="a8d08280-1def-11de-be11-b8a03c50a862", year="2013", limit=2000) | |
| library(RJSONIO) | |