Live data from Hacker News

Show HN: LeanReviews - quick reviews for almost anything

leanreviews.com

11–20 of 24 posts

Re: Show HN: LeanReviews - quick reviews for almost anything

#11
post #2

Hi HN, I am developing LeanReviews with a goal to make it easier to share opinions about anything people care about. People often have lots of opinions about lots of things but dont usually write elaborate reviews. LeanReviews makes sharing quick reviews almost trivial. I had submitted this to HN about a week ago but could not get any response. So I revamped some things and am resubmitting. I hope I get some feedback…

great job, but a favicon would be nice :)

Hey thanks for your feedback. We do need a logo and a favicon. Any suggestions for getting a logo designed?

Re: Show HN: LeanReviews - quick reviews for almost anything

#12
post #11

Earlier quoted context omitted.

great job, but a favicon would be nice :)

Hey thanks for your feedback. We do need a logo and a favicon. Any suggestions for getting a logo designed?

Budget: fiverr Professional: 99designs

Re: Show HN: LeanReviews - quick reviews for almost anything

#14
post #9
post #6

Earlier quoted context omitted.

you can also try separating negative and positive words in the word-cloud. Looks good!

Hey thanks for the suggestion. I was also thinking of doing something along those lines. However, dont have enough content presently. Some kind of sentiment analysis would be useful.

What kind of analysis are you doing so far with it?

Re: Show HN: LeanReviews - quick reviews for almost anything

#15
post #14
post #9

Earlier quoted context omitted.

Hey thanks for the suggestion. I was also thinking of doing something along those lines. However, dont have enough content presently. Some kind of sentiment analysis would be useful.

What kind of analysis are you doing so far with it?

Not much presently. Its just restricted to the bar chart. Infact that is mostly another form of visualisation of the content. There isnt much content presently for doing sentiment analysis where you could show a goodness meter for something based on the words being submitted as reviews. I am not finding much time to work on new features between collecting seed content, finding users, getting feedback, etc. Of course, that cannot be an excuse for not getting the features done. Infact there are lots of other items in the pipeline that need to be done - eg. spell check, more general search - presently its mostly exact match with some help from autocomplete, social integration, chat, etc.

Re: Show HN: LeanReviews - quick reviews for almost anything

#17
post #9
post #6

Earlier quoted context omitted.

you can also try separating negative and positive words in the word-cloud. Looks good!

Hey thanks for the suggestion. I was also thinking of doing something along those lines. However, dont have enough content presently. Some kind of sentiment analysis would be useful.

Could you use twitter to grab some reviews? You wouldn't need to do it ongoing but it might be a good way to get to a critical mass of reviews.

Re: Show HN: LeanReviews - quick reviews for almost anything

#18
post #17
post #9

Earlier quoted context omitted.

Hey thanks for the suggestion. I was also thinking of doing something along those lines. However, dont have enough content presently. Some kind of sentiment analysis would be useful.

Could you use twitter to grab some reviews? You wouldn't need to do it ongoing but it might be a good way to get to a critical mass of reviews.

Thats a great idea. Twitter can be a good source of seed content. Thanks for suggesting.

Re: Show HN: LeanReviews - quick reviews for almost anything

#20
Hey, great idea! I found a couple bugs:

1. There is not much user input validation, if any. For example, I submitted Broken to the San Francisco page, and it broke the "Explore Review Frequencies" section.

2. There is no checking for repeating submissions. Using a while loop on "send_review('stuff')" repeated my submission until I refreshed the page. A way to fix this would be to save an array of user ids per topic, but unfortunately that would require user login. Or, you could save an array of ip addresses, though obviously there are easy ways to get around that.

Post reply on HN