tidyquant 1.0.1.9000 (Development Version)

tidyquant 1.0.1

Improvements

  • tq_get() - Add "dividends" and "splits" get options, which have been fixed in quantmod 0.4-16. Issue 150.

Bug Fixes

  • Issue 157 - Error on package load with rstudioapi::getThemeInfo() returns NULL.
  • pivot_table - Fix issues with tidyverse functions not being found.

Deprecation

tidyquant 1.0.0

This is the “R for Excel Users” release. My aim is to build functionality that helps users coming from an Excel Background (background I came from). It’s important to have these users feel at home. I have a full suite of functionality to accomplish your Excel-to-R transition.

-Matt

  • Excel Functions

    • Why Excel functions? Designed to help users coming from an Excel Background more easily transition to the tidyverse and “tidy- finance / business analysis” in R.
    • Pivot Table
      • pivot_table() - A tidyverse-style function to perform data summarizations just like the popular Excel Pivot Table. Enables stacking calculations using a tidy-esque syntax: .rows = ~ YEAR(order_date).
    • Reference Functions
      • VLOOKUP() - Performs the classic Excel VLOOKUP. Excel user’s: rejoice.
    • Summarising “IFS” Functions
      • Summarising “IFS” Functions - Filtering versions of Excel summarization counterparts. Simply add “cases” that filter if a condition is true. SUM_IFS(), COUNT_IFS(), AVERAGE_IFS()
      • Create your own “IFS” functions - Have and idea for a new “IFS” function that hasn’t been made yet? Use CREATE_IFS() to make your own by supplying a summarization function.
    • Statistical, Date, and Financial Math Functions
  • NEW Tidyverse Functionality

    • summarise_by_time() - This is a new time-based variant of summarise() that allows collapsing the time-series by “day”, “week”, “month”, “quarter”, “year”, and more.
    • Note: I will evaluate the need for summarise_at_by_time(), summarise_all_by_time(), and summarise_if_by_time() after the release of dplyr v1.0.0.
  • NEW API Integrations

    • Tiingo API - A popular Open-Source for stock prices, cryptocurrencies, and intraday feeds from the IEX (Investors Exchange). This can serve as an alternate source of data to Yahoo Finance. Integrated via the riingo package.
  • Bug Fixes & Improvements

    • theme_tq() - Fix issues with collisions with dials::margin() and ggplot2::margin(). Similar potential ggplot2 collisions have been fixed.
    • theme_tq() - Increased default top/bottom text margin on facet strips
  • Deprecation & Breaking Changes

    • Potential Breaking Change - Single values now return the symbol column (i.e. tq_get("AAPL") returns symbol = “AAPL” for the 1st column).
    • Deprecated Sources: The following sources have been deprecated due to lack of support from the API:
      • Google Finance
      • Morningstar Key Ratios & Financials (Fundamentals) Data
      • Yahoo Dividends and Splits
      • Oanda FX and Metal Prices
    • Deprecated Compound Getters - Stacking multiple get options (tq_get("AAPL", get = c("stock.prices", "stock.prices.japan"))) is no longer available. Solution: Split these up into two calls to tq_get().

tidyquant 0.5.10

tidyquant 0.5.9

  • tq_index()
    • Fix issue #144 - tq_index() download issue. Note that “RUSSEL1000”, “RUSSELL2000”, “RUSSELL3000”, and “SP1000” are no longer available due to changes from www.us.spdrs.com.
    • Update Stock Index Fallback.

tidyquant 0.5.8

  • tq_index() - Fix naming issue with stock index data downloaded from www.us.spdrs.com.

tidyquant 0.5.7

Stock Index & Exchanges

Visualizations & Color Palettes

  • geom_candlestick and geom_barchart - Issue #112.
  • Added color names of theme_tq palettes (palette_light, palette_dark, and palette_green) for easier identification.

Compatability with tidyr v1.0.0

  • Improvements to ensure compatability with tidyr v1.0.0

[Potential Breaking Change] Move tidyverse to suggests

  • This is actually potentially a “breaking change” (although most users will see no difference since you likely load tidyverse in your scripts) - if you do not load tidyverse, then you will now need to do so. Previously tidyquant loaded tidyverse behind the scenes.

tidyquant 0.5.6

  • Morningstar Key Ratios: The tq_get() argument get = "key.ratios" has been deprecated due to a change in Morningstar’s website. (Help Wanted - Ref. Issue #125)

  • Remove dependency on XLConnect. Replace with readxl. Issue #119.

tidyquant 0.5.5

  • Bux fix

    • tq_get() get = "financials" now returns a warning and NA as Google Finance no longer provides data. We are actively looking for alternative data sources.

    • tq_get() get = "stock.prices.google" now returns a warning and NA as Google Finance no longer provides data. Use get = "stock.prices" instead to use Yahoo Finance, or use the riingo package to download from Tiingo.

    • Catch duplicate names in col_rename when you are renaming more than 1 column. Duplicate names are not allowed and return an error.

    • Fix duplicate name collision issue when the original name already includes a .. Duplicate names now get a ..1, ..2, etc. as opposed to .1, .2.

tidyquant 0.5.4

  • Features:
  • Important Changes:
    • Remove Key Statistics from tq_get(get = "key.stats"). Yahoo Finance no longer supports the Key Statistics CSV API.
    • Completed deprecation of tidyquant::as_tibble() and tidyquant::as_xts(). Use timetk::tk_tbl() and timetk::tk_xts() instead.
    • tibbletime support was added so that all tidyquant functions play nicely with tbl_time objects.
    • A hard dependency on XLConnect was removed. This should ease the use of the package, especially for Mac users.
  • Bug Fixes:
    • Some tests failed with testthat 2.0. They have been updated.

tidyquant 0.5.3

  • Fixes for compatibility with purrr v0.2.3.

tidyquant 0.5.2

  • Incorporated more robust timetk coercion functions. Deprecated tidyquant::as_xts() and tidyquant::as_tibble(). Use timetk::tk_xts() and timetk::tk_tbl() instead.
  • Fixes:
    • tq_index() no longer pulls from marketvolume. Instead, 9 indices are available from SPDR. These indices are more reliable, and include weights for each stock in the index.
    • Fixed 2 tests where the results of tq_get(get = "stock.prices") were 1 or 2 rows off of what the tests expected. This likely has to do with the new yahoo finance API.

tidyquant 0.5.1

  • Improvements
    • Added pkgdown integration.
  • Fixes:
    • Require new quantmod version 0.4-8 to fix Oanda and Yahoo bugs.
    • Quandl data returned newest to oldest. For consistency with other tq_get() data, it now returns oldest to newest.
    • Oanda only returns 180 days of FX and Metals data now. Updated the tests to account for this. Also added error handling to check for valid date ranges.
    • Fixed bug with tq_portfolio() where weights = NULL would not execute an equal weighting scheme.
    • Added error handling during dollar and percent conversion for get = “key.ratios” and get = “key.stats”.

tidyquant 0.5.0

tidyquant 0.4.0

  • New Features:
    • tq_transmute() replaces tq_transform() for consistency with dplyr.
    • tq_performance() which integrates the performance analysis functions of PerformanceAnalytics.
    • tq_portfolio() which enables aggregating portfolios from individual stock returns.
    • tq_tranform(): Added the NA-handling functions from zoo to the list of compatible, which provide a number of useful methods for handling NA values in data sets. Added Return.calculate and Return.excess for calculating returns and returns in excess of the risk-free rate, respectively.
  • Documentation:
    • tq_mutate() and tq_transmute() help pages have been combined.
    • Split introduction into four separate vignettes, which improves flow and enables readers to more easily get to needed documentation. Now five docs total covering the primary needs of tidyquant users!
  • New data:
  • New visualizations that integrate with ggplot2:
    • palette_() functions used to create scales are exported.
    • theme_tq() creates light, dark, and green themes for tidyquant visualizations.
    • scale_color_tq() and scale_fill_tq() add color/fill scales for the data used in tidyquant visualizations.
  • Improvements and Fixes:
    • The transform_fun argument of tq_transmute() has been replaced with mutate_fun for consistency with tq_mutate().
    • Core functions are now generics to allow for extendability.
    • Issue #11: Part 2. Fix multiple stocks that only return 110 lines. Handle stocks that return csv with “We’re sorry” message.
    • Issue #11: Part 1. Fix instability with get = key.ratios failing with HTTP 500 error on download. Use httr RETRY in case of failure.
    • Fixed issue with get = "key.ratios" where stocks listed on AMEX exchange were not able to return key ratios.
    • Issue #9: Fix problem with get = "key.stats" where NA’s in multiple x (e.g. c("AAPL", "GOOG")) cause call to fail during coercion.
    • Issue #8, Part 2: Enable compound gets (e.g. tq_get("AAPL", get = c("stock.prices", "financials"))).
    • Issue #8, Part 1: Create tq_index() function to return a stock index. tq_get(get = "stock.index") is deprecated and will be removed during the next version after 0.4.0. Use tq_index_options() for index options.
    • Issue #7: Fixed issue with date column inadvertently being coerced to dttm.

tidyquant 0.3.0

  • New data:
    • New tq_get option get = "key.stats", which retrieves the current key statistics (55 total) from www.finance.yahoo.com/. These include various current data such as Ask, Bid, Day’s High, Day’s Low, Last Trade Price, current P/E Ratio, EPS, Current Market Cap, EPS Projected Current Year, EPS Projected Next Year and many more. Example: tq_get("AAPL", get = "key.stats").
  • New visualizations that integrate with ggplot2:
    • Chart geoms: Bar charts (geom_barchart) and candlestick charts (geom_candlestick) can be quickly created with the new geoms.
    • Moving Averages: Seven moving averages can be quickly visualized / prototyped using geom_ma. The geom wraps the TTR::SMA functions.
    • Bollinger bands can be visualized with geom_bbands. The same seven moving averages are compatible with the geom.
    • Zooming Into Chart Sections: Two functions (coord_x_date and coord_x_datetime) were added to enable zooming into chart sections using dates with no out-of-bounds data loss (e.g. out-of-bounds data loss with the scale_x_ functions).
  • New Vignette: Covers “Charting with tidyquant”.
  • Fixes:
    • Issue #5: tq_get can now accept character vectors and data frames for the x arg, in addition to a single character input. This streamlines the getting of data for multiple inputs (e.g. stock symbols, stock indexes, etc).
    • Issue #4: Added col_rename arg to tq_mutate and tq_transform, which enables fast and easy renaming during the operation.
    • Issue #3: Integrated dplyr::group_by() with tq_mutate() and tq_transform(). The transform and mutate functions now work properly with grouped data frames.
    • Issue #2: Fixed bug with tq_get(), get = "key.ratios", where key ratios for stocks from the NYSE returned NA.
  • Removed support for deprecated arguments: x_fun, .x, and .y in the respective transform and mutate functions.

tidyquant 0.2.0

tidyquant 0.1.0

  • Initial release of tidyquant, for seamless quantitative financial analysis (xts, quantmod, TTR) package integration with the tidyverse.