Combined and shiny documentation

R/combmodels.R

 cmplot <-ggplot(cmdf, aes(x = x, y = y)) + geom_line(aes(y = f), color = "blue", size = 2) + geom_point() +
     geom_point(aes(y = r), color = "darkgreen")
 
-cmplot
+# cmplot
 
-cmplot + coord_cartesian(xlim = c(as.POSIXct("2017-03-01", tz = "UTC"), as.POSIXct("2017-04-01", tz = "UTC")))
+# cmplot + coord_cartesian(xlim = c(as.POSIXct("2017-03-01", tz = "UTC"), as.POSIXct("2017-04-01", tz = "UTC")))
 
-# sres <- stepAIC(charmmod, scope = list(upper = nform.full, lower = nform.min),
-#                 direction = "both", steps = 300)
 
 
 newagg <- p$read_pickle("../data/1617-agg.pkl")
 predplot <-ggplot(predf, aes(x = x, y = y)) + geom_line(aes(y = f), color = "blue", size = 2) + geom_point() +
     geom_point(aes(y = r), color = "darkgreen")
 
-predplot
+# predplot
 
-predplot + coord_cartesian(xlim = c(as.POSIXct("2017-03-01", tz = "UTC"), as.POSIXct("2017-04-01", tz = "UTC")))
+# predplot + coord_cartesian(xlim = c(as.POSIXct("2017-03-01", tz = "UTC"), as.POSIXct("2017-04-01", tz = "UTC")))
 
 mean(abs(predf$r))
 sd(predf$r)
 
 
 # number of icps per cluster
-ocdf <- p$read_pickle('../data/9-clusters-sample-table.pkl')
-ncdf <- p$read_pickle('../data/1617-asgn-table.pkl')
-table(ocdf$cluster)
-table(ncdf$cluster)
+# ocdf <- p$read_pickle('../data/9-clusters-sample-table.pkl')
+# ncdf <- p$read_pickle('../data/1617-asgn-table.pkl')
+# table(ocdf$cluster)
+# table(ncdf$cluster)
 
 
 allmods <- list()

README.md

 
 ### `weathmod.R`
 
-This script takes the downloaded weather data, (hard-coded path: `../data/2016-18-weather.pkl`) and calculates running last-24-hour minimum and maximum temperatures. It then fits harmonic regressions using the first and second harmonic to de-trend the yearly pattern of each series, saving the resulting dataset at `../data/weatherharm.csv` and stores the models in `../models/weatherMinharmonic.rds` and `../models/weatherMaxharmonic.rds`. These paths can be changed in the source, but there are no command line options.
+This script takes the downloaded weather data, (hard-coded path: `../data/2016-18-weather.pkl`) and calculates running last-24-hour minimum and maximum temperatures. It then fits harmonic regressions using the first and second harmonic to de-trend the yearly pattern of each series, saving the resulting dataset at `../data/weatherharm.csv` and stores the models in `../models/weatherMinharmonic.rds` and `../models/weatherMaxharmonic.rds`. These paths can be changed in the source (and will likely need to be as the data files are not included in the repository), but there are no command line options.
 
 Commented out in the source are a number of potentially informative plots showing this process. For example:
 
 
 ### `combmodels.R`
 
-This script takes the 
+Taking `../data/9-clusters.agg.pkl` and `../data/weatherharm.csv` as inputs, this script first fits a harmonic model to cluster 1 and tests it with the assigned clusters in `../data/1617-agg.pkl`. It then fits models to all clusters and saves them collectively in `../models/1kmods.rds`.
+
+Similar to the previous, some plots are again commented out.
+
+## Shiny App
+
+The [R Shiny](https://shiny.rstudio.com/) app is located in `shiny/app.R`. It requires the outputs from `weathmod.R` and `combmodels.R`, along with a couple of extra packages, installed with:
+
+```R
+install.packages(c("shiny", "shinycssloaders"))
+```
+
+Explaining Shiny and how to properly install it on a server in order to have a nice web app on a domain behind https _etc_ would take more time than I have (email me!) but for an example of a simple, self-hosted app in action see [this project](https://shiny.petras.space/Emissions/) I made awhile ago ([source](https://git.petras.space/petra/Emissions)).
+
+The app in this project can be run with `Rscript app.R`; this will (after several seconds of start-up) launch the app at a random port (e.g. 5576). Navigating to this location in your browser (firewall permitting) should display the app.
+
+The "data and model" tab allows exploration of the data along with the model projection; plots can be saved by right-clicking. The "prediction" tab allows, after specification of the number of icps in each of nine clusters, the temperature range, and the month and day of the week, the projected demand pattern. The month is approximate because it models a "typical" day in the approximate middle of that month, without a specific date.