Analyzing covariation of variables over time in R - r

Question

I am using Rstudio to analyze data. And - why else would I post this - I'm getting stuck!
The data is a time-series of the evolution of a community and their characteristics over time. For example, the n_members (the size) can increase over time.
Each combination of the variables 'randomseed' and 'lhsExperimentNumber' is 1 observation. In total, there are 80,000 observations. Each observation has a time-series going from 1 (month) to 204 (months). This time is characterized in the variable X.run.number
What I want:
See covariation of dependent and independent variables over time.
For example: I would like to see how the n_members (size) evolves over time, under different circumstances (exit_window, for example). I normally use dplyr and ggplot2, and I'd like to make some captivating images in ggplot.
My problem:
- I'm just a beginner in R so it's a little tricky for me to see the correct strategy here
- It would be sort of easier if I would just have 1 observation that evolves over time, but now I have 80,000 observations, which each independently evolve over time. So I don't really know how to work the data to make sensible graphs.
How would you guys tackle this? Maybe you can help me out with some tips (that would be so great!)
I used dput to reproduce the first 10 observations of my dataset:
library(ggplot2)
library(dplyr)
coop_dat <- structure(list(X.run.number. = c(2L, 3L, 4L, 1L, 2L, 3L, 4L,
1L, 3L, 3L), rewire_prop = c(0.1, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1,
0.1, 0.1, 0.1), prior_trust_std = c(0.1, 0.1, 0.1, 0.1, 0.1,
0.1, 0.1, 0.1, 0.1, 0.1), max_trust = c(2L, 2L, 2L, 2L, 2L, 2L,
2L, 2L, 2L, 2L), maxCR_trust = c(0.2, 0.2, 0.2, 0.2, 0.2, 0.2,
0.2, 0.2, 0.2, 0.2), Ostrom. = structure(c(2L, 2L, 2L, 2L, 2L,
2L, 2L, 2L, 2L, 2L), .Label = c("false", "true"), class = "factor"),
social_network = structure(c(1L, 1L, 1L, 1L, 1L, 1L, 1L,
1L, 1L, 1L), .Label = "\"small-world\"", class = "factor"),
external_ROI_influence = structure(c(2L, 2L, 2L, 2L, 2L,
2L, 2L, 2L, 2L, 2L), .Label = c("false", "true"), class = "factor"),
randomseed = c(1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L), lhsExperimentNumber = c(1L,
2L, 3L, 0L, 1L, 2L, 3L, 0L, 2L, 2L), X.step. = c(0L, 0L,
0L, 0L, 1L, 1L, 1L, 1L, 2L, 3L), size_start_cooperative = c(12L,
4L, 20L, 17L, 12L, 4L, 20L, 17L, 4L, 4L), degree_prior = c(7L,
21L, 40L, 15L, 7L, 21L, 40L, 15L, 21L, 21L), degree_coop = c(17L,
47L, 20L, 31L, 17L, 47L, 20L, 31L, 47L, 47L), prior_trust_average = c(0.40574414878618,
0.631650829650462, 0.369752783104777, 0.585182260912843,
0.40574414878618, 0.631650829650462, 0.369752783104777, 0.585182260912843,
0.631650829650462, 0.631650829650462), number_of_meetings = c(10L,
5L, 2L, 2L, 10L, 5L, 2L, 2L, 5L, 5L), n_interactions = c(20L,
3L, 9L, 19L, 20L, 3L, 9L, 19L, 3L, 3L), exit_window = c(5L,
1L, 3L, 4L, 5L, 1L, 3L, 4L, 1L, 1L), info_peer_behavior = c(0.396114811487496,
0.627111892865505, 0.148686024846975, 0.281594149058219,
0.396114811487496, 0.627111892865505, 0.148686024846975,
0.281594149058219, 0.627111892865505, 0.627111892865505),
memory = c(8L, 11L, 20L, 1L, 8L, 11L, 20L, 1L, 11L, 11L),
n_shares_total_sum = c(0L, 0L, 0L, 0L, 21L, 13L, 32L, 31L,
13L, 13L), n_shares_total_mean = c(0, 0, 0, 0, 1.75, 3.25,
1.6, 1.82352941176471, 3.25, 3.25), members_at_meeting = c(0L,
0L, 0L, 0L, 0L, 0L, 0L, 0L, 2L, 0L), trust_coop_total = c(0,
0, 0, 0, 0.0457418505048186, 0.0563109517340759, 0.0349050771003757,
0.0587220121136852, 0.0563109517340759, 0.0563109517340759
), rep_total = c(0, 0, 0, 0, 0, 0, 0, 0, 0, 0), ROI = c(0L,
0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L), current_strategy = c(0L,
0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L), price = c(0L, 0L, 0L,
0L, 0L, 0L, 0L, 0L, 0L, 0L), REfocus = c(0, 0, 0, 0, 0, 0,
0, 0, 0, 0), endpoint = c(0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L,
0L, 0L), HG_membertypes = c(0, 0, 0, 0, 0, 0, 0, 0, 0, 0),
n_members = c(0L, 0L, 0L, 0L, 12L, 4L, 20L, 17L, 4L, 4L),
n_ids = c(0L, 0L, 0L, 0L, 12L, 4L, 20L, 17L, 4L, 4L), n_prods = c(0L,
0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L), n_cons = c(0L, 0L, 0L,
0L, 0L, 0L, 0L, 0L, 0L, 0L), n_members_exit = c(0L, 0L, 0L,
0L, 0L, 0L, 0L, 0L, 0L, 0L)), .Names = c("X.run.number.",
"rewire_prop", "prior_trust_std", "max_trust", "maxCR_trust",
"Ostrom.", "social_network", "external_ROI_influence", "randomseed",
"lhsExperimentNumber", "X.step.", "size_start_cooperative", "degree_prior",
"degree_coop", "prior_trust_average", "number_of_meetings", "n_interactions",
"exit_window", "info_peer_behavior", "memory", "n_shares_total_sum",
"n_shares_total_mean", "members_at_meeting", "trust_coop_total",
"rep_total", "ROI", "current_strategy", "price", "REfocus", "endpoint",
"HG_membertypes", "n_members", "n_ids", "n_prods", "n_cons",
"n_members_exit"), row.names = c(NA, 10L), class = "data.frame")
EDIT:
Right now I have this simple ggplot code:
ggplot(dat) +
geom_line(aes(x=month, y=trust_coop_total))
This produces this graph:
You can see that this is not really a line graph. It's more a graph with a bunch of vertical lines. What I would like is to decipher trends for each individual sample. Each sample is, as mentioned above, a combination of lhsExperimentnumber and X.run.number. There are so many samples that a line graph would be overwhelming, but it might be nice in combination with the alpha=1/10 command of geom_jitter. However, geom_jitter ignores the trend in the sample itself.
What should I do to get a good trend graph?