v0.4.0

This is a minor release from 0.3.6 that includes bugfixes, performance improvements, and build changes.

Bug Fixes

#334
Pandas 1.0 fix: #364

New Features

Turnover tearsheet improvement: #354
CI builds now run on GitHub Actions: #363

Performance + API Change

#361 simplified the cumulative returns calculation, making it much faster. Other function signatures and behaviors were modified as well.

Docs + Miscellaneous

#332
README updates: #345, #337
A new tutorial notebook.

Credits

The following people contributed to this release:
@eigenfoo, @luca-s, @twiecki, @ivigamberdiev, @fawce, @jbredeche, @jimportico, @gerrymanoim, @jmccorriston, @altquant

v0.3.6

Add option to compute forward returns non-cumulatively

v0.3.5

This is a minor release from 0.3.4 that includes bugfixes, speed enhancement and compatibility with more recent pandas versions. We recommend that all users upgrade to this version.

Bugfixes

• Issue 323 factor_rank_autocorrelation infers turnover period in calendar space while periods could have different time space
• PR 324 avoid crashing Alphalens when autocorrelation or turnover data contains only NaNs

Performance

• PR 327 Speed up compute_forward_returns and get_clean_factor

Compatibility with new pandas versions

• PR 328 improved compatibility with pandas 0.23.4

v0.3.4

This is a minor release from 0.3.3 that includes bugfixes, small enhancements and backward compatibility breakages. We recommend that all users upgrade to this version.

Bugfixes

• PR 317 Fix date conversion in newer versions of pandas
• Issue 309 Biased Mean Quantile Returns for Non-Equal Bins

New features

• PR 306 added zero aware quantiles option

API change

• PR 268 All functions deprecated in version v0.2.0 are no longer available

Credits

The following people contributed to this release:

@eigenfoo - George Ho
@MichaelJMath - Mike Matthews
@freddiev4 - Freddie Vargus
@vikram-narayan - Vikram Narayan
@twiecki - Thomas Wiecki
@luca-s - Luca Scarabello

v0.3.2

This is a minor release from 0.3.1 that includes bugfixes and small enhancements. We recommend that all users upgrade to this version.

Bugfixes

• PR297 BUG: create_pyfolio_input doesn't work with frequency higher than 1 day
• PR302 BUG: compute_mean_return_spread returns error if no std_err argument

New features

• PR298 ENH: added rate of return option in 'create_event_study_tear_sheet'
• PR300 ENH: added n_bars option in 'create_event_study_tear_sheet'

v0.3.1

This is a minor release from 0.3.0 that includes bugfixes and performance improvement. We recommend that all users upgrade to this version.

Bugfixes

• PR 287 utils.get_clean_factor crashes with malformed 'groupby' data
• PR 287 perf.average_cumulative_return_by_quantile crahes in certain scenarios
• PR 288 monthly IC heatmap plot has inverted colors (red for positive and blue for negative IC)
• PR 295 Issue 292 utils.compute_forward_returns fails to detect the correct period length

Performance

• PR 294 computation of cumulative returns is very slow

v0.3.0

This is a major release from 0.2.1, we recommend that all users upgrade to this version.

New features

• Integration with Pyfolio. It is now possible to simulate a portfolio using the input alpha factor and analyze the performance with Pyfolio. Please see the relevant example notebook. [PR 227] and [PR 250]
• Added new API utils.get_clean_factor to run Alphalens with returns instead of prices [PR 270]
• Changed color palette to improve the visual experience for colorblind users [PR 248]
• Standard deviation bars optional in tears.create_event_returns_tear_sheet
• Alphalens now properly handles intraday factors

Bugfixes

• Alphalens now works with both tz-aware and tz-naive data (but not mixed)
• "Cumulative Returns by Quantile" plot used a different color scheme for quantiles than "Average Cumulative Returns by Quantile" plot
• Many small but useful bug fixes that avoid sporadic crashes and memory leaks. Please see the git history for more details

Documentation

• Added several new example Notebooks

Maintenance

• Removed deprecated pandas.TimeGrouper
• Migrated tests from deprecated nose-parameterized (#251)
• Fixed compatibility with matplotlib 2.2.0
• Alphalens is now available via conda-forge. Install via conda install -c conda-forge alphalens

Credits

The following people contributed to this release:

@luca-s - Luca Scarabello
@twiecki - Thomas Wiecki
@mmargenot - Max Margenot
@MichaelJMath
@HereticSK
@TimShawver - Tim Shawver
@alen12345 - Alessio Nava

v0.2.1

This is a bugfix release from v0.2.0. All users are recommended to upgrade.

Bugfixes

• tears.create_information_tear_sheet: argument group_adjust was erroneously removed without a replacement. From this release argumentgroup_adjust is still deprecated but group_neutral can be used instead

v0.2.0

This is a major new release since v0.1.0. It contains small API breakage, several new features and many bug fixes. All users are recommended to upgrade.

New since v0.1.0

New features

• Added event study analysis: an event study is a statistical method to assess the impact of a particular event on the value of equities and it is now possible to perform this analysis through the API alphalens.tears.create_event_study_tear_sheet. Check out the relative NoteBook in the example folder.

• Added support for group neutral factor analysis (group_neutral argument): this affects the return analysis that is now able to compute returns statistics for each group independently and aggregate them together assuming a portfolio where each group has equal weight.

• utils.get_clean_factor_and_forward_returns has a new parameter max_loss that controls how much data the function is allowed to drop due to not having enough price data or due to binning errors (pandas.qcut). This gives the users more control on what is happening and also avoid the function to raise an exception if the binning doesn't go well on some values.

• Greatly improved API documentation

Bugfixes

• Fix alpha and beta calculation on a Long only factor

• Inconsistent "Returns Analysis" Table

• Standard Error Conversion from n-Periods to 1-Period

• Improved help message for 'Bin edges must be unique' error and explain possible solutions

• ValueError: Invalid RGBA argument: 0.0`

• Cumulative returns plots not properly computed

• Update IC Risk Adjusted Ratio Calculation

API change

• Removed deprecated alphalens.tears.create_factor_tear_sheet
• tears.create_summary_tear_sheet: added argument group_neutral.
• tears.create_returns_tear_sheet: added argument group_neutral. Please consider using keyword arguments to avoid API breakage
• tears.create_information_tear_sheet: group_adjust is now deprecated and group_neutral should be used instead
• tears.create_full_tear_sheet: group_adjust is now deprecated and group_neutral should be used instead
• tears.create_event_returns_tear_sheet: added argument group_neutral. Please consider using keyword arguments to avoid API breakage
• Several small changes to lower level API (alphalens.performance)

Maintenance

• Depends on pandas>=0.18.0
• Changed deprecated pd.rolling_mean() to use the new *.rolling().mean() API
• Changed deprecated pd.rolling_apply() to use the new *.rolling().apply() API
• Use versioneer to pull version from git tag

v0.1.2

New release v0.1.2

• Removed deprecated API 'alphalens.tears.create_factor_tear_sheet'

• Added event study API 'alphalens.tears.create_event_study_tear_sheet' and relative example NB

• Added Long only option to 'alphalens.performance.factor_alpha_beta'

• Improved docstrings all around

• Small bug fixes