R-Ladies NYC: An Analysis

class: center, middle, inverse, title-slide

# R-Ladies NYC: An Analysis
## <br/>NYR 2020
### Ludmila Janda
### Amplify
### 2020/08/14

---

# WHAT IS R-LADIES?

[R-Ladies Global](https://benubah.github.io/r-community-explorer/rladies.html)

---

# WHAT IS THIS TALK?

- An exploration at the R-Ladies NYC meetup data

- A look at the code used to do this

---

# Getting data using meetupr

[meetupr](https://github.com/rladies/meetupr) github link

```r
urlname <- "rladies-newyork"
events <- meetupr::get_events(urlname, "past")
```

```r
colnames(events)
```

```
##  [1] "id"              "name"            "created"         "status"         
##  [5] "time"            "local_date"      "local_time"      "waitlist_count" 
##  [9] "yes_rsvp_count"  "venue_id"        "venue_name"      "venue_lat"      
## [13] "venue_lon"       "venue_address_1" "venue_city"      "venue_state"    
## [17] "venue_zip"       "venue_country"   "description"     "link"           
## [21] "resource"
```

```r
nrow(events)
```

```
## [1] 47
```

---

# Are our RSVP counts trending upwards?

```r
summary(events$yes_rsvp_count)
```

```
##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##   20.00   42.00   55.00   56.68   71.50  103.00
```

```r
ggplot(events, aes(local_date, yes_rsvp_count)) + 
  geom_line(color = "#88398a") +
  geom_point(color = "#88398a", size = 2) + 
  labs(x = "Date", y = "RSVP Count") + 
  theme_rladies()
```

---

# Are our RSVP counts trending upwards?

---

# What are those low RSVP count events?

```r
events_low_labels <- events %>% 
  mutate(name = ifelse(yes_rsvp_count > 35, "", name), 
         name = str_remove(name, "R-ladies"), 
         name = str_remove(name, "R-Ladies"),
         name = str_remove(name, "R Ladies"))

ggplot(events_low_labels, 
       aes(local_date, yes_rsvp_count, label = name)) + 
  geom_line(color = "#88398a") +
  geom_point(color = "#88398a", size = 2) + 
  geom_label_repel(size = 2.5, 
                   force = 5, 
                   nudge_y = 5, 
                   segment.color = "grey60") + 
  labs(x = NULL, y = "RSVP Count") + 
  theme_rladies()
```

---

# What are those low RSVP count events?

---

# Filter out expectedly small events

```r
events_normal <- events %>% 
  filter(!str_detect(name, "Book Club"), 
         !str_detect(name, "book club"), 
         !str_detect(name, "Social"), 
         !str_detect(name, "Workshop")) 
```

```r
summary(events_normal$yes_rsvp_count)
```

```
##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##   38.00   55.00   66.00   68.72   82.00  103.00
```

---

# Trending Up?

```r
ggplot(events_normal, aes(local_date, yes_rsvp_count)) + 
  geom_line(color = "#88398a") +
  geom_point(color = "#88398a", size = 2) + 
  labs(x = NULL, y = "RSVP Count") + 
  theme_rladies()
```

---

# Trending Up?

---

# What were those "popular" events?

```r
events_normal_labels <- events_normal %>% 
  mutate(name = ifelse(yes_rsvp_count < 80, "", name))

ggplot(events_normal_labels, 
       aes(local_date, yes_rsvp_count, label = name)) + 
  geom_col(fill = "#88398a") +
  geom_label_repel(size = 4, 
                   force = 5, 
                   nudge_y = -7, 
                   segment.color = "grey60") + 
  labs(x = NULL, y = "RSVP Count") + 
  theme_rladies()
```

---

# What were those "popular" events?

---

# Look at members data

```r
members <- meetupr::get_members(urlname)
```

```r
table(members$status)
```

```
## 
## active 
##   1856
```

---

# Count of new members by month

```r
members_month <- members %>% 
  mutate(year_month = as.Date(floor_date(created, "month"))) %>% 
  group_by(year_month) %>% 
  summarise(count_new_members = n())

ggplot(members_month, aes(year_month, count_new_members)) + 
  geom_col(fill = "#88398a") + 
  scale_x_date(date_breaks = '6 months', date_labels = '%Y-%m') +
  labs(x = NULL, y = "Number of New Members") +
  theme_rladies() 
```

---

# Count of new members by month

---

# New members by nearest upcoming event

```r
# group by day and get count of new accounts
# use complete() to get all the days
new_members <- members %>% 
  mutate(date = as.Date(created)) %>% 
  group_by(date) %>% 
  summarise(count_new_members = n()) %>% 
  tidyr::complete(date = seq.Date(min(date), max(date), by = "day"))

# make event dataset to merge with
events_dates <- events %>% 
  select(date = local_date, name) 
```

---

# New members by nearest upcoming event

```r
# merge members with complete dates and events 
# get counts by event
members_events_count <- 
  left_join(new_members, events_dates, by = "date") %>%
  mutate(event_flag = ifelse(!is.na(name), 1, 0), 
         event_group = cumsum(event_flag) - event_flag) %>% 
  group_by(event_group) %>% 
  summarise(new_members = sum(count_new_members, na.rm = TRUE))

# get event dates for event groups
members_events_groups <- 
  left_join(new_members, events_dates, by = "date") %>% 
  mutate(event_flag = ifelse(!is.na(name), 1, 0), 
         event_group = cumsum(event_flag) - event_flag) %>% 
  filter(!is.na(name)) %>% 
  distinct(event_group, name, date)

# final join
members_events <- 
  left_join(members_events_groups, members_events_count, 
            by = "event_group") %>% 
  mutate(name = ifelse(new_members < 50, "", name))
```

---

# New members by nearest upcoming event

```r
members_events %>% 
  mutate(name = 
           str_remove(name, "R-Ladies New York Lecture Series: ")) %>% 
  ggplot(aes(date, new_members, label = name)) + 
  geom_col(color = "black") + 
  geom_label_repel(size = 4, 
                   force = 40, 
                   nudge_y = 40, 
                   segment.color = "grey60") + 
  labs(x = NULL, y = "New Member Count") +
  theme_rladies() 
```

---

# New members by nearest upcoming event

---

# Hadley Wickham Bump?

```r
members_events <- left_join(members_events_groups, 
                            members_events_count, 
                            by = "event_group") %>% 
  mutate(name = 
           ifelse(str_detect(name, "Tidy evaluation"), 
                  name, ""))

ggplot(members_events, 
       aes(date, new_members, label = name)) + 
  geom_col(color = "black") + 
  geom_label_repel(size = 4, 
                   force = 7, 
                   nudge_y = 10, 
                   segment.color = "grey60") + 
  labs(x = NULL, y = "New Member Count") +
  theme_rladies() 
```

---

# Hadley Wickham Bump?

---

# Member counts with events

```r
members_events_all_days <- 
  left_join(new_members, events_dates, by = "date") %>% 
  mutate(event_flag = 
           factor(ifelse(!is.na(name), "Yes", "No")), 
         count_new_members = 
           ifelse(is.na(count_new_members), 0, 
                  count_new_members), 
         year = year(date))

ggplot(members_events_all_days, 
       aes(date, count_new_members, 
           fill = event_flag, color = event_flag)) + 
  geom_col() + 
  facet_wrap(~year, scales = "free") +
  scale_fill_manual(values = c("#D3D3D3", "#88398a")) +
  scale_color_manual(values = c("#D3D3D3", "#88398a"), guide = FALSE) +
  scale_x_date(date_breaks = "2 months", date_labels = "%m") +
  labs(x = "Month", y = "New Member Count", 
       fill = "Event Day") +
  theme_rladies() 
```

---

# Member counts with events

---

# Loyalty

```r
# all attendees for one event
attendees <- 
  meetupr::get_event_attendees(urlname, 
                               event_id = "235073330")

# get all event ids 
event_ids <- events$id

# get attendees for all the events
all_event_attendees <- 
  event_ids %>% 
  map_dfr(slowly(~(get_event_attendees(urlname, event_id = .)))) 
```

---

# Loyalty

```r
attendee_rsvp_counts <- 
  all_event_attendees %>% 
  filter(id != 0, 
         rsvp_response == "yes") %>% 
  group_by(id) %>% 
  summarise(count_events_rsvp = n()) %>% 
  left_join(members, by = "id") %>% 
  filter(created < "2020-01-01")

ggplot(attendee_rsvp_counts, aes(count_events_rsvp)) +
  geom_histogram(binwidth = 1, 
                 fill = "#88398a") +
  labs(x = "Number of Events Attended", y = "Count") +
  theme_rladies()
```

---

# Loyalty

---

# R-Ladies NYC Number One Fan

```r
subset(attendee_rsvp_counts, 
       id == 103401142, 
       select = count_events_rsvp)
```

```
## # A tibble: 1 x 1
##   count_events_rsvp
##               <int>
## 1                40
```

---

# Map of Event Locations

```r
shps <- here::here("nyc")
boroughs <- st_read(shps, "nyc") %>% 
  filter(boro_name == "Manhattan" | boro_name == "Brooklyn") %>% 
  pull(geometry)
```

```
## Reading layer `nyc' from data source `/Users/ljanda/Documents/misc/nyr_talk_2020/nyc' using driver `ESRI Shapefile'
## Simple feature collection with 4 features and 4 fields
## geometry type:  MULTIPOLYGON
## dimension:      XY
## bbox:           xmin: -74.04773 ymin: 40.54183 xmax: -73.70001 ymax: 40.91553
## epsg (SRID):    4326
## proj4string:    +proj=longlat +datum=WGS84 +no_defs
```

```r
nyc_boroughs = st_transform(boroughs, 2263)

events_spatial <- 
  events %>% 
  select(latitude = venue_lat, longitude = venue_lon) %>% 
  filter(!is.na(latitude)) %>% 
  st_as_sf(coords = c("longitude", "latitude"), crs = 4326) %>% 
  st_transform(2263)
```

---

# Map of Event Locations

```r
ggplot() + 
  geom_sf(data = nyc_boroughs, fill = "white", color = "darkgrey") +
  geom_sf(data = events_spatial, 
          alpha = 0.25, size = 2, color = "#88398a") +  
  labs(x = "", y = "") +
  coord_sf() +
  theme_minimal() +
  theme(panel.border = element_blank(),
        panel.grid.major = element_blank(),
        panel.grid.minor = element_blank(),
        axis.line = element_line(colour = "white"),
        axis.text.x = element_blank(),
        axis.text.y = element_blank(),
        axis.ticks = element_blank())
```

---
background-image: url("images/event_locations.png")
background-size: 125%
background-position: center top

# Event Locations

---

# Further Work

- flexdashboard

- automate website event posts

- consider who our speakers are/if we are involving the community

---

# Word Cloud - Comments

---

# Thank You

- Sebastian Teran Hidalgo

- Laura Janda

- Samuel Crane

- Ayanthi Gunawardana

- R-Ladies NYC!!!