i have ggplot2 chart looks like:
the dataframe looks like:
city weekday time avg_wait_time timestamp 1511 boston tuesday 09:06 0.20 2001-01-04 09:06:00 1512 boston tuesday 09:07 0.20 2001-01-04 09:07:00 1513 boston tuesday 09:08 0.20 2001-01-04 09:08:00 1514 boston tuesday 09:09 3.17 2001-01-04 09:09:00 1515 boston tuesday 09:10 3.17 2001-01-04 09:10:00 1516 boston tuesday 09:11 3.17 2001-01-04 09:11:00 1517 boston tuesday 09:12 3.17 2001-01-04 09:12:00 1518 boston tuesday 09:13 5.87 2001-01-04 09:13:00 1519 boston tuesday 09:14 5.87 2001-01-04 09:14:00 1520 boston tuesday 09:15 5.87 2001-01-04 09:15:00 1521 boston tuesday 09:16 5.87 2001-01-04 09:16:00 1522 boston tuesday 09:17 9.17 2001-01-04 09:17:00 1523 boston tuesday 09:18 9.17 2001-01-04 09:18:00 1524 boston tuesday 09:19 12.20 2001-01-04 09:19:00 1525 boston tuesday 09:20 12.20 2001-01-04 09:20:00 1526 boston tuesday 09:21 12.10 2001-01-04 09:21:00 1527 boston tuesday 09:23 13.70 2001-01-04 09:23:00 1528 boston tuesday 09:24 13.70 2001-01-04 09:24:00 1529 boston tuesday 09:25 15.30 2001-01-04 09:25:00 1530 boston tuesday 09:26 15.30 2001-01-04 09:26:00 1531 boston tuesday 09:27 16.90 2001-01-04 09:27:00 1532 boston tuesday 09:28 16.90 2001-01-04 09:28:00 1533 boston tuesday 09:29 18.33 2001-01-04 09:29:00 1534 boston tuesday 09:30 18.33 2001-01-04 09:30:00 1535 boston tuesday 09:31 16.90 2001-01-04 09:31:00 1536 boston tuesday 09:32 16.90 2001-01-04 09:32:00 1537 boston tuesday 09:33 18.57 2001-01-04 09:33:00 1538 boston tuesday 09:34 18.57 2001-01-04 09:34:00 1539 boston tuesday 09:35 21.73 2001-01-04 09:35:00 1540 boston tuesday 09:36 21.73 2001-01-04 09:36:00 the classes dataframe are:
> sapply(x_output, class) $city [1] "factor" $weekday [1] "factor" $time [1] "character" $avg_wait_time [1] "numeric" $timestamp [1] "posixct" "posixt" as can see graphs, there whitespace because ggplot2 plotting 24 hour day. in dataset, data 1 business-week time period (2001-01-01 2001-01-05) (an "average" week these specific dates arbitrarily set) , business hours between 9am , 6pm.
how can make ggplot2 plot between 9am , 6pm exclusively?
here's attempt:
p <- ggplot(x_output, aes(x=timestamp, y=avg_wait_time, group=city)) + geom_line(aes(color=city), size=1.5) + theme(axis.text.x = element_text(angle = 90, hjust=1), legend.position = "bottom") + labs(x=null, y="waiting time (minutes)") + facet_wrap( ~ weekday, ncol=5) + scale_x_datetime(breaks = date_breaks("1 hour")), limits = c(as.posixct("9:00"), as.posixct("18:00")) print(p) i think if can configure limits part correctly, ggplot2 plot properly. help.
update: got step closer using (see scales part):
p <- ggplot(x_output, aes(x=timestamp, y=avg_wait_time, group=city)) + geom_line(aes(color=city), size=1.5) + theme(axis.text.x = element_text(angle = 90, hjust=1), legend.position = "bottom") + labs(x=null, y="waiting time (minutes)") + facet_wrap( ~ weekday, ncol=5, **scales="free"**) + scale_x_datetime(breaks = date_breaks("1 hour"), labels=date_format("%h:%m")) now output looks like:
however, can tell image, axis wrong , doesn't reflect data. data looks (i've put copy here: http://pastebin.com/sdhzkhpc):
> r <- r[order(r$timestamp),] > head(r,15) city weekday time avg_wait_time timestamp 2714 boston thursday 10:04 0.00 2001-01-06 10:04:00 2715 boston thursday 10:05 0.00 2001-01-06 10:05:00 2716 boston thursday 10:06 0.23 2001-01-06 10:06:00 2717 boston thursday 10:07 0.23 2001-01-06 10:07:00 2718 boston thursday 10:08 3.33 2001-01-06 10:08:00 2719 boston thursday 10:09 3.33 2001-01-06 10:09:00 2720 boston thursday 10:10 4.80 2001-01-06 10:10:00 2721 boston thursday 10:11 4.80 2001-01-06 10:11:00 2722 boston thursday 10:12 6.33 2001-01-06 10:12:00 2723 boston thursday 10:13 6.33 2001-01-06 10:13:00 2724 boston thursday 10:14 7.90 2001-01-06 10:14:00 2725 boston thursday 10:15 7.90 2001-01-06 10:15:00 2726 boston thursday 10:16 9.50 2001-01-06 10:16:00 2727 boston thursday 10:17 9.50 2001-01-06 10:17:00 2728 boston thursday 10:18 12.17 2001-01-06 10:18:00 > tail(r,15) city weekday time avg_wait_time timestamp 2699 boston thursday 05:41 pm 22.07 2001-01-06 17:41:00 2700 boston thursday 05:42 pm 23.47 2001-01-06 17:42:00 2701 boston thursday 05:43 pm 23.47 2001-01-06 17:43:00 2702 boston thursday 05:44 pm 24.90 2001-01-06 17:44:00 2703 boston thursday 05:45 pm 24.90 2001-01-06 17:45:00 2704 boston thursday 05:46 pm 26.43 2001-01-06 17:46:00 2705 boston thursday 05:47 pm 19.33 2001-01-06 17:47:00 2706 boston thursday 05:49 pm 12.23 2001-01-06 17:49:00 2707 boston thursday 05:50 pm 0.00 2001-01-06 17:50:00 2708 boston thursday 05:52 pm 0.23 2001-01-06 17:52:00 2709 boston thursday 05:54 pm 1.80 2001-01-06 17:54:00 2710 boston thursday 05:55 pm 1.80 2001-01-06 17:55:00 2711 boston thursday 05:56 pm 0.00 2001-01-06 17:56:00 2712 boston thursday 05:57 pm 0.00 2001-01-06 17:57:00 2713 boston thursday 05:58 pm 0.00 2001-01-06 17:58:00 update 2: timezone issue.
i solved using resource: plotting times ggplot: added hour in plot maybe due daylight saving?
more specifically, did:
tz(x_output$timestamp) <- "gmt" # "america/new_york" the comment included trial test several different time zones see matched (they can found here: http://www.inside-r.org/packages/cran/lubridate/docs/tz).
i keep question posterity.
my approach not use timestamp x axis rather use time.
x_output$time <- strftime(timestamp, format="%h:%m") then setting ggplot using call same weekday facet should work.
p <- ggplot(x_output, aes(x=time, y=avg_wait_time, group=city)) nb: don't have access r environment right cannot directly verify if code works in principle should :)
Comments
Post a Comment