first cut at input data figure for poster #50
Conversation
|
What version of EGRET do you have? I have (from EGRET * 2.6.1.9003 2017-11-22 Github (USGS-R/EGRET@b941b9f) |
|
EGRET 2.6.0 2016-07-27 CRAN (R 3.4.4) |
|
Ah, better: |
|
oops, yeah had the wrong function call |
9_report/src/fig_input_data.R
Outdated
| ref_date = '2017-05-09' | ||
|
|
||
| preds <- readRDS(sc_retrieve(preds_ind, remake_file)) %>% | ||
| dplyr::filter(ref_date == '2017-05-09', # dplyr::filter doesn't recognize objects?? |
There was a problem hiding this comment.
objects? what kinds of object did you want it to recognize?
There was a problem hiding this comment.
the ones defined on lines 4-6 (ref_date, etc...). Or could replace with a yml file as input.
There was a problem hiding this comment.
dplyr by default uses non-standard evaluation (hence we can use ref_date and it knows we're talking about the column). However, there's been a lot of work recently to make it friendlier to standard evaluation. This should work:
preds <- readRDS(sc_retrieve(preds_ind, remake_file)) %>%
dplyr::filter(ref_date == !!(ref_date ),
Basically, how would it know when you are talking about column names vs variable(object)? Surround it by !! (which...apparently is said out loud "bang-bang")
There was a problem hiding this comment.
ah thanks! bang-bang!
!!(thanks)
There was a problem hiding this comment.
woot! I forgot about !! and would have suggested just renaming ref_date. @ldecicco-USGS 's suggestion is way better.
9_report/src/fig_input_data.R
Outdated
| Site == '05465500') | ||
|
|
||
| # example of input data for a forecast | ||
| eList <- prep_inputs(nwis_site = site, |
| theme(axis.title.x = element_blank())) | ||
|
|
||
| #concentration | ||
| g2 <- ggplotGrob(ggplot(data = eList$Sample, aes(x = Date, y = ConcAve)) + |
There was a problem hiding this comment.
we could think about using lollipops or randomized residuals for censored data, a la https://owi.usgs.gov/blog/rResid/. @ldecicco-USGS, want to weigh in on this?
There was a problem hiding this comment.
I don't have a favorite, but it would be a good touch to flag censored data. Could do a simple color a la https://owi.usgs.gov/blog/plotFlowConc/
You could also just use the EGRET plot functions. multiPlotDataOverview is a nice intro plot to the data.
Merge branch 'master' of git://github.com/USGS-R/protoloads # Conflicts: # 3_forecast/src/prep_inputs.R
3_forecast/src/prep_inputs.R
Outdated
| # period for this "analysis" substitute | ||
| nwm_analysis <- bind_rows( | ||
| nwm_retro %>% filter(date >= as.Date('2016-11-08'), date < min(nwm_forecast$valid_date)), | ||
| nwm_forecast %>% filter(ref_date == valid_date) %>% select(site_no, date = valid_date, flow)) |
There was a problem hiding this comment.
the deletions in this file look wrong to me - i think i just added this and want to keep it
There was a problem hiding this comment.
yeah I had some merge conflicts that I think I screwed up but I'm pushing up your changes now
interactive = FALSEand it was happy. Not sure if I have a different version of EGRET or not...