code examples/r-gas/gas_model.R
In ahardisty/dstrain:
gc()
train_test <- fread('../CEA_DATA/weather_data/train_test.txt')
hdd_dt <- fread('../CEA_DATA/weather_data/hdd_dt.txt')
baseline_usage <- fread('../CEA_DATA/weather_data/baseline_usage.txt')
# Fit and predict -----------------------------------------------
gc()
random_customers <- train_test %>%
distinct(customer) %>%
sample_frac(1) %>%
ungroup()
# start_time <- proc.time()
train <- train_test%>%
filter(customer %in% (random_customers$customer)) %>%
filter(train_indicator == 'TRAIN') %>%
filter(model_name == 'TEMP_NORMAL') %>%
filter(outlier == FALSE) %>% # keep outliers out of modeling
filter(climate_zone_cd %in% CLIMATE_ZONES) %>% # keep only climate zones of interest
group_by(customer, prem_id, acct_id, opr_area_cd, climate_zone_cd, train_indicator) %>%
nest(.key = TRAIN) %>%
# spread(key = train_indicator, value = data)%>%
mutate(
models = map(TRAIN, usg_temp_lm), # fit linear model
augment_train = map(models, broom::augment),
glance_train = map(models, broom::glance),
tidy_train = map(models, broom::tidy)#model level
)
test <- train_test %>%
filter(customer %in% (random_customers$customer)) %>%
filter(train_indicator == 'TEST') %>%
filter(model_name == 'TEMP_NORMAL') %>%
filter(outlier == FALSE) %>% # keep outliers out of modeling
filter(climate_zone_cd %in% CLIMATE_ZONES) %>% # keep only climate zones of interest
group_by(customer, prem_id, acct_id, opr_area_cd, climate_zone_cd, train_indicator) %>%
nest(.key = TEST)
test <- test %>%
mutate(
.pred = map2(train$models, TEST, predict) # make out of sample predictions
)
# Calculate model fit -----------------------------------------------------
# train %>% select(glance_train) %>% unnest()
# train %>% select(tidy_train) %>% unnest()
# train %>% select(models) %>% map(1)
# head(train)\1
# extract elements from train list in data frame d
train %>% select(tidy_train) %>% unnest()
train %>% select(glance_train) %>% unnest()
train_df <- train %>%
select(customer, prem_id, acct_id, opr_area_cd, climate_zone_cd, train_indicator,
TRAIN, augment_train) %>%
mutate(
.pred = augment_train %>% map(".fitted"), #fitted values
.resid = augment_train %>% map('.resid'),
usg = TRAIN %>% map('usg'),
usage_actual = TRAIN %>% map('usage_actual'), #usage values
month_name = TRAIN %>% map('month_name'),
year = TRAIN %>% map('year'), #date values
year_day = TRAIN %>% map('year_day'), #date values
temp_normal = TRAIN %>% map('temp_normal'), #date values
temp_actual = TRAIN %>% map('temp_actual'), #date values
model_name = TRAIN %>% map('model_name'),
month_day = TRAIN %>% map('month_day'),
month = TRAIN %>% map('month'),
date = TRAIN %>% map('date'),
year = TRAIN %>% map('year'), #date values
season = TRAIN %>% map('season'),
sa_id = TRAIN %>% map('sa_id') #usage values
# opr_area_cd = TRAIN %>% map('opr_area_cd'), #usage values
# season = TRAIN %>% map('season'), #usage values
# train_indicator = 'TRAIN'
) %>%
select(-c(TRAIN, augment_train)) %>%
unnest()
train_summary <- train_df %>%
group_by(customer, model_name, climate_zone_cd, month_name, train_indicator) %>%
summarise(y = sum(usage_actual)
, y_hat = sum(.pred)
, resid = sum(.resid)) %>%
group_by(model_name, climate_zone_cd, month_name, train_indicator) %>%
summarize(rmse = rmse(y_hat = y_hat, y = y), n = n()) %>%
gather(key = variable, value = model_value, c(rmse)) %>%
ungroup() %>%
as.data.frame()
test_df <- test %>%
select(customer, prem_id, acct_id, opr_area_cd, climate_zone_cd, train_indicator, TEST, .pred) %>%
mutate(
usage_actual = TEST %>% map('usage_actual'), #usage values
usage_prior = TEST %>% map('usage_prior'), #usage values
model_name = TEST %>% map('model_name'),
usg = TEST %>% map('usg'),
year_day = TEST %>% map('year_day'), #date values
temp_normal = TEST %>% map('temp_normal'), #date values
temp_actual = TEST %>% map('temp_actual'), #date values
month_name = TEST %>% map('month_name'),
month_day = TEST %>% map('month_day'),
month = TEST %>% map('month'),
date = TEST %>% map('date'),
year = TEST %>% map('year'), #date values
season = TEST %>% map('season'),
sa_id = TEST %>% map('sa_id')
# climate_zone_cd = TEST %>% map('climate_zone_cd'), #usage values
# opr_area_cd = TEST %>% map('opr_area_cd'), #usage values
# train_indicator = 'TEST'
) %>%
select(-c(TEST)) %>%
unnest() %>%
mutate(.resid = usage_actual - .pred)
test_summary <- test_df %>%
group_by(customer, model_name, climate_zone_cd, month_name, train_indicator) %>%
summarise(y = sum(usage_actual)
, y_hat = sum(.pred)
, resid = sum(.resid)) %>%
group_by(model_name, climate_zone_cd, month_name, train_indicator) %>%
summarize(rmse = rmse(y_hat = y_hat, y = y), n = n()) %>%
gather(key = variable, value = model_value, c(rmse)) %>%
ungroup() %>%
as.data.frame()
baseline <- baseline_usage %>%
# rename(.pred = pred) %>%
filter(outlier == FALSE) %>% # keep outliers out of modeling
filter(climate_zone_cd %in% CLIMATE_ZONES) %>% # keep only climate zones of interest
filter(customer %in% (random_customers$customer)) %>%
mutate(.resid = usage_actual - .pred) %>%
tbl_df()
baseline_summary <- baseline %>%
group_by(customer, model_name, climate_zone_cd, month_name, train_indicator) %>%
summarise(y = sum(usage_actual)
, y_hat = sum(.pred)
, resid = sum(.resid)) %>%
group_by(model_name, climate_zone_cd, month_name, train_indicator) %>%
summarize(rmse = rmse(y_hat = y_hat, y = y), n = n()) %>%
gather(key = variable, value = model_value, c(rmse)) %>%
ungroup() %>%
as.data.frame()
# combine in sample, out of sample and baseline model details
model_comp <- rbind(test_summary, train_summary, baseline_summary) %>%
mutate(model_name = paste0(model_name,'_',train_indicator)) %>%
arrange(desc(month_name), model_name) %>%
left_join(baseline_summary %>% select(baseline_model = model_name, baseline = model_value, climate_zone_cd, month_name)
, by = c('climate_zone_cd','month_name'))
model_comp <- model_lift(model_comp, baseline = baseline, model_value = model_value )
model_comp <- model_comp %>%
select(model_name, climate_zone_cd, n, month_name, train_indicator, variable, model_value, baseline:model_lift) %>%
filter(model_name %in% c('TEMP_NORMAL_TEST','CONS_USG_BASELINE'))
# Plotting model fit ------------------------------------------------------
# format dates for plotting
model_comp$model_name <- as.factor(model_comp$model_name)
model_comp$month_name <- factor(model_comp$month_name, levels = c('Nov','Dec','Jan','Feb','Mar'))
# summary for plots
plot_details <- model_comp %>%
filter(train_indicator == 'TEST') %>%
select(climate_zone_cd, n) %>%
group_by(climate_zone_cd, n) %>%
distinct() %>%
ungroup %>%
mutate(total = sum(n))
# how do the models compare to baseline on average during the winter?
rmse_comparision <- ggplot(data = model_comp %>%
filter(train_indicator!='TRAIN' & model_name %in%
c('CONS_USG_BASELINE','TEMP_NORMAL_TEST')) %>%
group_by(model_name, climate_zone_cd, month_name, train_indicator) %>%
summarize(average_model_pct_change = round(mean(model_pct,2)),
average_model_num_change = round(mean(model_delta),2),
average_model_value = round(mean(model_value),2),
average_baseline_num = mean(baseline)), aes(x = month_name, y = average_model_value, fill = model_name)) +
geom_bar(stat = 'identity', position = 'dodge') +
facet_wrap(~climate_zone_cd, ncol = 1) +
geom_text(aes(label = average_model_value), position = position_dodge(1)) +
scale_fill_discrete("Comparing Accuracy \nAcross Model Types"
,label = c('Constant Usage'
, 'Temperature Model'))+
ggtitle(paste("Comparing model RMSE for gas usage for",unique(plot_details$total),'customers in'
,'T&W')) +
labs(y = 'RMSE \n(lower is better)', x = NULL)
ggsave(plot = rmse_comparision, filename = 'model_rmse_comparison.png'
, path = 'images/', width = 11, height = 8.5)
# percent improvement of temperature compared to baseline (constant usage)
rmse_pct <- ggplot(model_comp %>% filter(model_name %in% c('TEMP_NORMAL_TEST')),
aes(x = month_name, y = model_pct, fill = model_name)) +
geom_bar(stat = 'identity', aes(y=model_pct, ymin=0, ymax=model_pct), position = 'dodge')+
geom_text(aes(x = month_name, y = model_pct, ymax = model_pct, label = model_pct,
vjust = ifelse(sign(model_pct)>0, 1, -1)), position = position_dodge(width = 1))+
facet_wrap(~climate_zone_cd, ncol = 1)+
scale_fill_discrete("Model Type",
labels = 'Temperature Model') +
ggtitle("Comparing percent improvement in RMSE for\nnormal temperature over Constant Usage") +
labs(y = 'percent improvement (+/- over baseline)', x = NULL)
ggsave(plot = rmse_pct, filename = 'model_rmse_pct_improvement.png', path = 'images/', width = 11, height = 8.5)
# Plotting HDD comparisons ---------------------------------------------------------
# comparing actual and normal HDD 65 for train and test periods
hdd_dt$month_name <- factor(hdd_dt$month_name, levels = c('Nov','Dec','Jan','Feb','Mar'))
hdd_comp <- ggplot(hdd_dt %>% filter(hdd_type != 'hdd_norm_2014'))+
geom_bar(aes(x =month_name, y = hdd_65, fill = hdd_type), stat = 'identity'
, position = 'dodge') +
facet_wrap(~climate_zone_cd, ncol = 1) +
scale_fill_discrete(name="Heating Degree Days"
,labels=c('2014 HDD-65 Actual',
'2015 HDD-65 Actual',
'2015 HDD-65 Normal')) +
labs(y = "Heating Degree Days", x = NULL)+
ggtitle("Comparing Actual and Normal HDD-65")
ggsave(plot = hdd_comp, filename = 'hdd_comparison.png', path = 'images/', width = 11, height = 8.5)
# Export files ------------------------------------------------------------
gc()
# need to map this
test_out <- test_df %>%
select(-usage_prior) %>%
filter(model_name == 'TEMP_NORMAL') %>%
select(train_indicator, customer, prem_id, acct_id, sa_id, opr_area_cd, climate_zone_cd, .pred, .resid
, usage_actual,model_name, usg, year_day, month_name, month_day, month, date, year, season)
train_out <- train_df %>%
filter(model_name == 'TEMP_NORMAL') %>%
select(train_indicator, customer, prem_id, acct_id, sa_id, opr_area_cd, climate_zone_cd, .pred, .resid
, usage_actual, model_name, usg, year_day, month_name, month_day, month, date, year, season)
baseline_out <- baseline %>%
mutate(usg = usage_actual) %>%
select(train_indicator, customer, prem_id, acct_id, sa_id, opr_area_cd, climate_zone_cd, .pred, .resid
, usage_actual,model_name, usg, year_day, month_name, month_day, month, date, year, season)
# need to walk this
write.table(x = test_out, file = '../CEA_DATA/model_predictions/test.txt'
, col.names = TRUE, row.names = FALSE)
write.table(x = train_out, file = '../CEA_DATA/model_predictions/train.txt'
, col.names = TRUE, row.names = FALSE)
write.table(x = baseline_out, file = '../CEA_DATA/model_predictions/baseline.txt'
, col.names = TRUE, row.names = FALSE)
#
# rm(train_test)
# rm(test)
# rm(train)
# rm(baseline_usage)
ahardisty/dstrain documentation built on May 28, 2019, 4:42 p.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
gc()
train_test <- fread('../CEA_DATA/weather_data/train_test.txt')
hdd_dt <- fread('../CEA_DATA/weather_data/hdd_dt.txt')
baseline_usage <- fread('../CEA_DATA/weather_data/baseline_usage.txt')
# Fit and predict -----------------------------------------------
gc()
random_customers <- train_test %>%
distinct(customer) %>%
sample_frac(1) %>%
ungroup()
# start_time <- proc.time()
train <- train_test%>%
filter(customer %in% (random_customers$customer)) %>%
filter(train_indicator == 'TRAIN') %>%
filter(model_name == 'TEMP_NORMAL') %>%
filter(outlier == FALSE) %>% # keep outliers out of modeling
filter(climate_zone_cd %in% CLIMATE_ZONES) %>% # keep only climate zones of interest
group_by(customer, prem_id, acct_id, opr_area_cd, climate_zone_cd, train_indicator) %>%
nest(.key = TRAIN) %>%
# spread(key = train_indicator, value = data)%>%
mutate(
models = map(TRAIN, usg_temp_lm), # fit linear model
augment_train = map(models, broom::augment),
glance_train = map(models, broom::glance),
tidy_train = map(models, broom::tidy)#model level
)
test <- train_test %>%
filter(customer %in% (random_customers$customer)) %>%
filter(train_indicator == 'TEST') %>%
filter(model_name == 'TEMP_NORMAL') %>%
filter(outlier == FALSE) %>% # keep outliers out of modeling
filter(climate_zone_cd %in% CLIMATE_ZONES) %>% # keep only climate zones of interest
group_by(customer, prem_id, acct_id, opr_area_cd, climate_zone_cd, train_indicator) %>%
nest(.key = TEST)
test <- test %>%
mutate(
.pred = map2(train$models, TEST, predict) # make out of sample predictions
)
# Calculate model fit -----------------------------------------------------
# train %>% select(glance_train) %>% unnest()
# train %>% select(tidy_train) %>% unnest()
# train %>% select(models) %>% map(1)
# head(train)\1
# extract elements from train list in data frame d
train %>% select(tidy_train) %>% unnest()
train %>% select(glance_train) %>% unnest()
train_df <- train %>%
select(customer, prem_id, acct_id, opr_area_cd, climate_zone_cd, train_indicator,
TRAIN, augment_train) %>%
mutate(
.pred = augment_train %>% map(".fitted"), #fitted values
.resid = augment_train %>% map('.resid'),
usg = TRAIN %>% map('usg'),
usage_actual = TRAIN %>% map('usage_actual'), #usage values
month_name = TRAIN %>% map('month_name'),
year = TRAIN %>% map('year'), #date values
year_day = TRAIN %>% map('year_day'), #date values
temp_normal = TRAIN %>% map('temp_normal'), #date values
temp_actual = TRAIN %>% map('temp_actual'), #date values
model_name = TRAIN %>% map('model_name'),
month_day = TRAIN %>% map('month_day'),
month = TRAIN %>% map('month'),
date = TRAIN %>% map('date'),
year = TRAIN %>% map('year'), #date values
season = TRAIN %>% map('season'),
sa_id = TRAIN %>% map('sa_id') #usage values
# opr_area_cd = TRAIN %>% map('opr_area_cd'), #usage values
# season = TRAIN %>% map('season'), #usage values
# train_indicator = 'TRAIN'
) %>%
select(-c(TRAIN, augment_train)) %>%
unnest()
train_summary <- train_df %>%
group_by(customer, model_name, climate_zone_cd, month_name, train_indicator) %>%
summarise(y = sum(usage_actual)
, y_hat = sum(.pred)
, resid = sum(.resid)) %>%
group_by(model_name, climate_zone_cd, month_name, train_indicator) %>%
summarize(rmse = rmse(y_hat = y_hat, y = y), n = n()) %>%
gather(key = variable, value = model_value, c(rmse)) %>%
ungroup() %>%
as.data.frame()
test_df <- test %>%
select(customer, prem_id, acct_id, opr_area_cd, climate_zone_cd, train_indicator, TEST, .pred) %>%
mutate(
usage_actual = TEST %>% map('usage_actual'), #usage values
usage_prior = TEST %>% map('usage_prior'), #usage values
model_name = TEST %>% map('model_name'),
usg = TEST %>% map('usg'),
year_day = TEST %>% map('year_day'), #date values
temp_normal = TEST %>% map('temp_normal'), #date values
temp_actual = TEST %>% map('temp_actual'), #date values
month_name = TEST %>% map('month_name'),
month_day = TEST %>% map('month_day'),
month = TEST %>% map('month'),
date = TEST %>% map('date'),
year = TEST %>% map('year'), #date values
season = TEST %>% map('season'),
sa_id = TEST %>% map('sa_id')
# climate_zone_cd = TEST %>% map('climate_zone_cd'), #usage values
# opr_area_cd = TEST %>% map('opr_area_cd'), #usage values
# train_indicator = 'TEST'
) %>%
select(-c(TEST)) %>%
unnest() %>%
mutate(.resid = usage_actual - .pred)
test_summary <- test_df %>%
group_by(customer, model_name, climate_zone_cd, month_name, train_indicator) %>%
summarise(y = sum(usage_actual)
, y_hat = sum(.pred)
, resid = sum(.resid)) %>%
group_by(model_name, climate_zone_cd, month_name, train_indicator) %>%
summarize(rmse = rmse(y_hat = y_hat, y = y), n = n()) %>%
gather(key = variable, value = model_value, c(rmse)) %>%
ungroup() %>%
as.data.frame()
baseline <- baseline_usage %>%
# rename(.pred = pred) %>%
filter(outlier == FALSE) %>% # keep outliers out of modeling
filter(climate_zone_cd %in% CLIMATE_ZONES) %>% # keep only climate zones of interest
filter(customer %in% (random_customers$customer)) %>%
mutate(.resid = usage_actual - .pred) %>%
tbl_df()
baseline_summary <- baseline %>%
group_by(customer, model_name, climate_zone_cd, month_name, train_indicator) %>%
summarise(y = sum(usage_actual)
, y_hat = sum(.pred)
, resid = sum(.resid)) %>%
group_by(model_name, climate_zone_cd, month_name, train_indicator) %>%
summarize(rmse = rmse(y_hat = y_hat, y = y), n = n()) %>%
gather(key = variable, value = model_value, c(rmse)) %>%
ungroup() %>%
as.data.frame()
# combine in sample, out of sample and baseline model details
model_comp <- rbind(test_summary, train_summary, baseline_summary) %>%
mutate(model_name = paste0(model_name,'_',train_indicator)) %>%
arrange(desc(month_name), model_name) %>%
left_join(baseline_summary %>% select(baseline_model = model_name, baseline = model_value, climate_zone_cd, month_name)
, by = c('climate_zone_cd','month_name'))
model_comp <- model_lift(model_comp, baseline = baseline, model_value = model_value )
model_comp <- model_comp %>%
select(model_name, climate_zone_cd, n, month_name, train_indicator, variable, model_value, baseline:model_lift) %>%
filter(model_name %in% c('TEMP_NORMAL_TEST','CONS_USG_BASELINE'))
# Plotting model fit ------------------------------------------------------
# format dates for plotting
model_comp$model_name <- as.factor(model_comp$model_name)
model_comp$month_name <- factor(model_comp$month_name, levels = c('Nov','Dec','Jan','Feb','Mar'))
# summary for plots
plot_details <- model_comp %>%
filter(train_indicator == 'TEST') %>%
select(climate_zone_cd, n) %>%
group_by(climate_zone_cd, n) %>%
distinct() %>%
ungroup %>%
mutate(total = sum(n))
# how do the models compare to baseline on average during the winter?
rmse_comparision <- ggplot(data = model_comp %>%
filter(train_indicator!='TRAIN' & model_name %in%
c('CONS_USG_BASELINE','TEMP_NORMAL_TEST')) %>%
group_by(model_name, climate_zone_cd, month_name, train_indicator) %>%
summarize(average_model_pct_change = round(mean(model_pct,2)),
average_model_num_change = round(mean(model_delta),2),
average_model_value = round(mean(model_value),2),
average_baseline_num = mean(baseline)), aes(x = month_name, y = average_model_value, fill = model_name)) +
geom_bar(stat = 'identity', position = 'dodge') +
facet_wrap(~climate_zone_cd, ncol = 1) +
geom_text(aes(label = average_model_value), position = position_dodge(1)) +
scale_fill_discrete("Comparing Accuracy \nAcross Model Types"
,label = c('Constant Usage'
, 'Temperature Model'))+
ggtitle(paste("Comparing model RMSE for gas usage for",unique(plot_details$total),'customers in'
,'T&W')) +
labs(y = 'RMSE \n(lower is better)', x = NULL)
ggsave(plot = rmse_comparision, filename = 'model_rmse_comparison.png'
, path = 'images/', width = 11, height = 8.5)
# percent improvement of temperature compared to baseline (constant usage)
rmse_pct <- ggplot(model_comp %>% filter(model_name %in% c('TEMP_NORMAL_TEST')),
aes(x = month_name, y = model_pct, fill = model_name)) +
geom_bar(stat = 'identity', aes(y=model_pct, ymin=0, ymax=model_pct), position = 'dodge')+
geom_text(aes(x = month_name, y = model_pct, ymax = model_pct, label = model_pct,
vjust = ifelse(sign(model_pct)>0, 1, -1)), position = position_dodge(width = 1))+
facet_wrap(~climate_zone_cd, ncol = 1)+
scale_fill_discrete("Model Type",
labels = 'Temperature Model') +
ggtitle("Comparing percent improvement in RMSE for\nnormal temperature over Constant Usage") +
labs(y = 'percent improvement (+/- over baseline)', x = NULL)
ggsave(plot = rmse_pct, filename = 'model_rmse_pct_improvement.png', path = 'images/', width = 11, height = 8.5)
# Plotting HDD comparisons ---------------------------------------------------------
# comparing actual and normal HDD 65 for train and test periods
hdd_dt$month_name <- factor(hdd_dt$month_name, levels = c('Nov','Dec','Jan','Feb','Mar'))
hdd_comp <- ggplot(hdd_dt %>% filter(hdd_type != 'hdd_norm_2014'))+
geom_bar(aes(x =month_name, y = hdd_65, fill = hdd_type), stat = 'identity'
, position = 'dodge') +
facet_wrap(~climate_zone_cd, ncol = 1) +
scale_fill_discrete(name="Heating Degree Days"
,labels=c('2014 HDD-65 Actual',
'2015 HDD-65 Actual',
'2015 HDD-65 Normal')) +
labs(y = "Heating Degree Days", x = NULL)+
ggtitle("Comparing Actual and Normal HDD-65")
ggsave(plot = hdd_comp, filename = 'hdd_comparison.png', path = 'images/', width = 11, height = 8.5)
# Export files ------------------------------------------------------------
gc()
# need to map this
test_out <- test_df %>%
select(-usage_prior) %>%
filter(model_name == 'TEMP_NORMAL') %>%
select(train_indicator, customer, prem_id, acct_id, sa_id, opr_area_cd, climate_zone_cd, .pred, .resid
, usage_actual,model_name, usg, year_day, month_name, month_day, month, date, year, season)
train_out <- train_df %>%
filter(model_name == 'TEMP_NORMAL') %>%
select(train_indicator, customer, prem_id, acct_id, sa_id, opr_area_cd, climate_zone_cd, .pred, .resid
, usage_actual, model_name, usg, year_day, month_name, month_day, month, date, year, season)
baseline_out <- baseline %>%
mutate(usg = usage_actual) %>%
select(train_indicator, customer, prem_id, acct_id, sa_id, opr_area_cd, climate_zone_cd, .pred, .resid
, usage_actual,model_name, usg, year_day, month_name, month_day, month, date, year, season)
# need to walk this
write.table(x = test_out, file = '../CEA_DATA/model_predictions/test.txt'
, col.names = TRUE, row.names = FALSE)
write.table(x = train_out, file = '../CEA_DATA/model_predictions/train.txt'
, col.names = TRUE, row.names = FALSE)
write.table(x = baseline_out, file = '../CEA_DATA/model_predictions/baseline.txt'
, col.names = TRUE, row.names = FALSE)
#
# rm(train_test)
# rm(test)
# rm(train)
# rm(baseline_usage)
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.