In this example, we’ll use a simple linear model to identify the trend in annual returns to determine if the stock returns are decreasing or increasing over time.
Analyze a Single Stock
First, let’s collect stock data with tq_get()
AAPL <- tq_get("AAPL", from = "2007-01-01", to = "2016-12-31")
AAPL
## # A tibble: 2,518 x 8
## symbol date open high low close volume adjusted
## <chr> <date> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
## 1 AAPL 2007-01-03 12.3 12.4 11.7 12.0 309579900 10.4
## 2 AAPL 2007-01-04 12.0 12.3 12.0 12.2 211815100 10.6
## 3 AAPL 2007-01-05 12.3 12.3 12.1 12.2 208685400 10.5
## 4 AAPL 2007-01-08 12.3 12.4 12.2 12.2 199276700 10.6
## 5 AAPL 2007-01-09 12.4 13.3 12.2 13.2 837324600 11.4
## 6 AAPL 2007-01-10 13.5 14.0 13.4 13.9 738220000 12.0
## 7 AAPL 2007-01-11 13.7 13.8 13.6 13.7 360063200 11.8
## 8 AAPL 2007-01-12 13.5 13.6 13.3 13.5 328172600 11.7
## 9 AAPL 2007-01-16 13.7 13.9 13.6 13.9 311019100 12.0
## 10 AAPL 2007-01-17 13.9 13.9 13.5 13.6 411565000 11.7
## # … with 2,508 more rows
Next, come up with a function to help us collect annual log returns. The function below mutates the stock prices to period returns using tq_transmute(). We add the type = "log" and period = "monthly" arguments to ensure we retrieve a tibble of monthly log returns. Last, we take the mean of the monthly returns to get MMLR.
get_annual_returns <- function(stock.returns) {
stock.returns %>%
tq_transmute(select = adjusted,
mutate_fun = periodReturn,
type = "log",
period = "yearly")
}
Let’s test get_annual_returns out. We now have the annual log returns over the past ten years.
AAPL_annual_log_returns <- get_annual_returns(AAPL)
AAPL_annual_log_returns
## # A tibble: 10 x 2
## date yearly.returns
## <date> <dbl>
## 1 2007-12-31 0.860
## 2 2008-12-31 -0.842
## 3 2009-12-31 0.904
## 4 2010-12-31 0.426
## 5 2011-12-30 0.228
## 6 2012-12-31 0.282
## 7 2013-12-31 0.0776
## 8 2014-12-31 0.341
## 9 2015-12-31 -0.0306
## 10 2016-12-30 0.118
Let’s visualize to identify trends. We can see from the linear trend line that AAPL’s stock returns are declining.
AAPL_annual_log_returns %>%
ggplot(aes(x = year(date), y = yearly.returns)) +
geom_hline(yintercept = 0, color = palette_light()[[1]]) +
geom_point(size = 2, color = palette_light()[[3]]) +
geom_line(size = 1, color = palette_light()[[3]]) +
geom_smooth(method = "lm", se = FALSE) +
labs(title = "AAPL: Visualizing Trends in Annual Returns",
x = "", y = "Annual Returns", color = "") +
theme_tq()
Now, we can get the linear model using the lm() function. However, there is one problem: the output is not “tidy”.
mod <- lm(yearly.returns ~ year(date), data = AAPL_annual_log_returns)
mod
##
## Call:
## lm(formula = yearly.returns ~ year(date), data = AAPL_annual_log_returns)
##
## Coefficients:
## (Intercept) year(date)
## 58.86279 -0.02915
We can utilize the broom package to get “tidy” data from the model. There’s three primary functions:
augment: adds columns to the original data such as predictions, residuals and cluster assignmentsglance: provides a one-row summary of model-level statisticstidy: summarizes a model’s statistical findings such as coefficients of a regression
We’ll use tidy to retrieve the model coefficients.
## # A tibble: 2 x 5
## term estimate std.error statistic p.value
## <chr> <dbl> <dbl> <dbl> <dbl>
## 1 (Intercept) 58.9 113. 0.520 0.617
## 2 year(date) -0.0291 0.0562 -0.518 0.618
Adding to our workflow, we have the following:
get_model <- function(stock_data) {
annual_returns <- get_annual_returns(stock_data)
mod <- lm(yearly.returns ~ year(date), data = annual_returns)
tidy(mod)
}
Testing it out on a single stock. We can see that the “term” that contains the direction of the trend (the slope) is “year(date)”. The interpetation is that as year increases one unit, the annual returns decrease by 3%.
## # A tibble: 2 x 5
## term estimate std.error statistic p.value
## <chr> <dbl> <dbl> <dbl> <dbl>
## 1 (Intercept) 58.9 113. 0.520 0.617
## 2 year(date) -0.0291 0.0562 -0.518 0.618
Now that we have identified the trend direction, it looks like we are ready to scale.
Scale to Many Stocks
Once the analysis for one stock is done scale to many stocks is simple. For brevity, we’ll randomly sample ten stocks from the S&P500 with a call to dplyr::sample_n().
set.seed(10)
stocks_tbl <- tq_index("SP500") %>%
sample_n(5)
stocks_tbl
## # A tibble: 5 x 8
## symbol company identifier sedol weight sector shares_held local_currency
## <chr> <chr> <chr> <chr> <dbl> <chr> <dbl> <chr>
## 1 FLS Flowserve … 34354P10 22884… 1.37e-4 Indus… 1363520 USD
## 2 EW Edwards Li… 28176E10 25671… 1.66e-3 Healt… 6646630 USD
## 3 MTB M&T Bank C… 55261F10 23401… 5.18e-4 Finan… 1409828 USD
## 4 WAT Waters Cor… 94184810 29376… 4.33e-4 Healt… 662973 USD
## 5 HWM Howmet Aer… 44320110 BKLJ8… 2.29e-4 Indus… 4132548 USD
We can now apply our analysis function to the stocks using dplyr::mutate and purrr::map. The mutate() function adds a column to our tibble, and the map() function maps our custom get_model function to our tibble of stocks using the symbol column. The tidyr::unnest function unrolls the nested data frame so all of the model statistics are accessable in the top data frame level. The filter, arrange and select steps just manipulate the data frame to isolate and arrange the data for our viewing.
stocks_model_stats <- stocks_tbl %>%
select(symbol, company) %>%
tq_get(from = "2007-01-01", to = "2016-12-31") %>%
# Nest
group_by(symbol, company) %>%
nest() %>%
# Apply the get_model() function to the new "nested" data column
mutate(model = map(data, get_model)) %>%
# Unnest and collect slope
unnest(model) %>%
filter(term == "year(date)") %>%
arrange(desc(estimate)) %>%
select(-term)
stocks_model_stats
## # A tibble: 4 x 7
## # Groups: symbol, company [4]
## symbol company data estimate std.error statistic p.value
## <chr> <chr> <list> <dbl> <dbl> <dbl> <dbl>
## 1 MTB M&T Bank Corporation <tibble [2,… 0.0460 0.0243 1.90 0.0945
## 2 EW Edwards Lifesciences… <tibble [2,… 0.00458 0.0368 0.124 0.904
## 3 WAT Waters Corporation <tibble [2,… 0.00121 0.0416 0.0292 0.977
## 4 FLS Flowserve Corporation <tibble [2,… -0.0372 0.0498 -0.746 0.477
We’re done! We now have the coefficient of the linear regression that tracks the direction of the trend line. We can easily extend this type of analysis to larger lists or stock indexes. For example, the entire S&P500 could be analyzed removing the sample_n() following the call to tq_index("SP500").
