Feedback by UserVoice

Brett Ables

My feedback

  1. 66 votes
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)

    We’ll send you updates on this idea

    This is great feedback. We are looking into the details of this. Please continue to vote on feature ideas which will increase their likelihood of getting in the product.

    Thanks,
    Scott [MSFT]

    Brett Ables commented  · 

    I'll add to this that "Data Exploration" is now a far more common practice than it used to be. It is very common to get data off the internet, plot it, and then want to explore the data in greater detail by zooming in and panning around similar to exploring a region on a map. This is especially true for dense datasets or time-based data. Daily data for 10 years is overwhelming, but the ability to zoom in to see how weekly trends evolved over time or track a single line among hundreds helps use discern patterns much more quickly as we come to understand the data. Coming from an engineering background, every plot we make is an X, Y scatter plot used to discern shapes or trends of related variables. Some features are evident from the default zoom level, but often zooming in is necessary to see interesting data points. It is also very common in Engineering to plot 2D drawings and shapes on an XY plot. In this scenario zooming is just as useful as in any drawing or CAD software. It's understandable if bar and pie graphs are seen as the primary use case why zoom/pan aren't available, but for line/scatter plots it is an essential feature. The lack of interactive plotting is the biggest reason why Engineers / Scientists / Researchers move to more powerful tools to do "real work."

    Brett Ables supported this idea  · 
    Brett Ables commented  · 

    Chart interactivity as a whole could be drastically improved with planning and zooming. Changing the axis bounds to explore data is ridiculously slow and not user friendly currently... which drives people to better plotting tools in Matlab, Python, etc if they need to interact with complex data.

  2. 115 votes
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)

    We’ll send you updates on this idea

    Brett Ables supported this idea  · 
  3. 435 votes
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)

    We’ll send you updates on this idea

    Brett Ables commented  · 

    Translating regex capabilities to Worksheet Functions does require a lot of consideration since regex can be used in many different workflows. Matching (True/False), Splitting, token extraction, searching, and counting are all various workflows that can utilize regex pattern matching. This would require a family of functions that implement different capabilities, with some aimed at user-friendly simple usage and a couple other array functions for power users. (Think slope, intercept, forecast vs the power-user function linear).

    The most obvious use to me would be to replace the existing workflow to extract something from a string which is currently a painful combination of find, left, right, mid, and substitute. By default the entire match is returned; if the pattern contains tokens, and optional argument allows selecting which of the tokens to return. A more powerful array function would be capable of returning multiple matches containing multiple tokens as a 2D array which could either be given a worksheet range to populate or aggregated to a single result with a function like Index() like any other array function.

    Similar functions could perform replacement, counting, splitting, etc. Each regex related function would accept text, a pattern, and the standard regex options (ignorecase, dotall, multiline, etc.) plus additional arguments specific to each type of operation.

    While most languages these days have a somewhat object oriented approach to Regex returning complex Match or MatchCollection objects (to use VBA's nomenclature), this would need to be a more functional implementation similar to what Matlab does with regexp and regexprep.

    Brett Ables supported this idea  · 
  4. 6,318 votes
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)

    We’ll send you updates on this idea

    Wow.

    Thank you to all those participated in our survey. What an amazing number of responses, many of them with very deep content. We’re processing the over ten thousand responses(!), and already appreciate the time so many of you took to answer with passion and experience.

    Please know this survey is used to help influence various topics – both on Python as well as other related topics that the comments started to bleed into. Given the passion, I want to be clear this remains an area of exploration for us, without any specific timeline.

    We’ll provide updates as we progress on this feature request.

    Thanks!

    Ashvini Sharma
    Lead Program Manager
    Excel

    Brett Ables supported this idea  · 
  5. 38 votes
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)

    We’ll send you updates on this idea

    Brett Ables commented  · 

    This is the kind of common data visualization task that drives me to other tools like Python, using pandas to collate/slice/group data, and seaborn/matplotlib to plot it. In seaborn, not only can the color be based on another column, but also plots can be faceted with rows and columns of plots where the rows and columns are also based on other columns.

    https://stanford.edu/~mwaskom/software/seaborn/tutorial/axis_grids.html

    These are also features available with R and ggplot (from which pandas and seaborn get much inspiration).

    Brett Ables supported this idea  · 
  6. 17 votes
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)

    We’ll send you updates on this idea

    Brett Ables commented  · 

    Scott,

    I certainly agree that many of the creative uses of Excel are to fill or work around gaps in Excel's capabilities!

    Regarding plotting geometric figures... I think plotting of equations is a fantastic idea that would save time and "worksheet scratch space" for examples like the one I showed. In my experience as an engineer at NASA I constantly see folks implementing geometries of physical parts in Excel for a variety of reasons. Sometimes these geometries are circles, arcs, ellipses, or other shapes that can be represented by equations, but often they are custom curves pulled from a drawing or CAD file. I think parametric equation plotting, or iso-curve plotting (think contour lines from an elevation map or isobars from a weather radar map), would be valuable time savers that enable visualizing a relationship without first generating "dummy data" from the relationship.

    However, this "Chart Feature: Dynamic Legends" feature request is addressing the specific limitation of Excel Charts' legends. At the end of the day, the relationship between the legend seen on the chart and the series that are plotted is very static. If I create a legend, it defaults to including all series in the legend. If I have lines and dots and they are color coordinated, I may only want legend entries for the lines to avoid clutter. To do this I have to then delete the legend entries for the dots, 1 by 1, by hand. If I then have a dynamic setting that groups the dots into dots, triangles and squares based on type, but maintains the color association with the lines, I now have more legend entries to delete. And if for some reason I want some of those legend entries back or I delete the wrong one, I believe I'd have to delete the entire legend, and start over. Excel encourages and enables the creation of dynamic plots, but is hampered by the static nature of the legends. Honestly the static nature of the axes bounds is even more annoying in this regard, but this feature is addressing the legends specifically.

    Brett Ables commented  · 

    I often create interactive sheets with options that may result in adding or removing curves from a plot. I can add or remove the curves from the plot easily with an if() statement and setting the series data to na() so that the series isn't plotted. However, the legend will still show these "invisible" series.

    Here's an example: https://www.dropbox.com/s/1icwvpucsu3qxph/LegendExample.xlsx?dl=0

    In the example I'm plotting pill-shaped figure (two semi-circles with a rectangular space between them), allowing the user to select the radius and middle length. If the middle length is 0, the figure simply becomes a circle and the extension is "deactivated" with formulas which set it's plotting data to na(). However, the legend still shows the entry for the series. This is a very simple example, where much more complex examples exaggerate this effect.

    I've made data plotting sheets allowing users to select up to 10 parameters to plot. Once plotted, the names of the parameters are shown in the legend. However, if the user has only selected 2 parameters, without macros I have no choice but to have 8 empty legend entries.

    There should be an option to hide legend entries for series that aren't visible because they are empty or na(), or perhaps an option to hide legends for series when nothing is visible in the current view. Example: A plot showing latitude and longitude locations of restaurants by chain could then automatically reduce the legend dynamically to only the 3 chains represented in the town being focused on in the plot.

    Another tactic would be to allow the legend to be dynamically removed if the series name is set to na(). This would at least allow programmatic control for dynamic legend display.

    Brett Ables supported this idea  · 
  7. 6 votes
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)

    We’ll send you updates on this idea

    Brett Ables commented  · 

    I could see this working in OneNote which already attempts to fill this role, but doesn't handle embedding as well as it should.

  8. 15 votes
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)

    We’ll send you updates on this idea

    Brett Ables supported this idea  · 
  9. 100 votes
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)

    We’ll send you updates on this idea

    Brett Ables supported this idea  · 
  10. 9 votes
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)

    We’ll send you updates on this idea

    Brett Ables supported this idea  · 
  11. 87 votes
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)

    We’ll send you updates on this idea

    Brett Ables supported this idea  · 
  12. 4 votes
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)

    We’ll send you updates on this idea

    Brett Ables supported this idea  · 
  13. 95 votes
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)

    We’ll send you updates on this idea

    Brett Ables supported this idea  · 
  14. 44 votes
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)

    We’ll send you updates on this idea

    Brett Ables supported this idea  · 
  15. 6 votes
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)

    We’ll send you updates on this idea

    Brett Ables supported this idea  · 
  16. 13 votes
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)

    We’ll send you updates on this idea

    Brett Ables supported this idea  · 
  17. 6 votes
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)

    We’ll send you updates on this idea

    1 comment  ·  Excel for Windows (Desktop Application) » Other  ·  Flag idea as inappropriate…  ·  Admin →
    Brett Ables supported this idea  · 
  18. 27 votes
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)

    We’ll send you updates on this idea

    Brett Ables supported this idea  · 
  19. 12 votes
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)

    We’ll send you updates on this idea

    Brett Ables supported this idea  · 
  20. 14 votes
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)

    We’ll send you updates on this idea

    Brett Ables commented  · 

    More advanced Interpolation, integration, regression, filtering.... take a look at Python's scipy or Matlab's capabilities...

    Brett Ables supported this idea  · 
← Previous 1 3 4

Feedback and Knowledge Base